Abstract In the Internet of Things, the CBAC (capability‐based access control) model is proposed to make devices securely accessed. Here, an owner of a device issues a capability token, that is, a set of access rights, to a subject. The subject is allowed to manipulate ROs (resource objects) in the device according to access rights in the capability token. In the CBAC model, there is a problem a subject sbi can obtain data of an RO by accessing the data in an RO after the data are brought to the RO even if the subject sbi is not allowed to obtain the data from the RO . Here, the data in the RO illegally flow to the subject sbi. In our previous studies, the OI (operation interruption) protocol where illegal operations are interrupted is proposed and evaluated in the simulation. In this article, we implement the OI protocol and evaluate the authorization process of the OI protocol in terms of the execution time. In the evaluation, we make clear the features of the execution time of authorization processes for GET, PUT, POST, and DELETE operations in the OI protocol.

Summary The Internet of Things is composed of sensor and actuator devices. Devices have to be securely accessed by subjects. In this article, we take the capability‐based access control (CBAC) model where a subject is issued a capability token to manipulate a device by a device owner. In the CBAC model, information which a subject is not allowed to get may illegally flow to the subject. Hence, the operation interruption (OI) protocol to prevent illegal information flow is proposed in our previous studies. However, although a subject is not allowed to get data at time , the subject can get the data later than the time . Here, the data come to the subject later than expected by the subject to get the data, that is, the information flows late to the subject. In this article, we newly propose a time‐based OI (TBOI) protocol to prevent not only illegal information flow but also late information flow. Here, operations implying illegal or late information flow are interrupted, that is, not performed at devices. In the evaluation, we show not only illegal information flow but also late information flow are prevented in the TBOI protocol differently from the OI protocol.

Summary In the fog computing (FC) models, a fog node supports not only routing functions but also application processes. By the application processes, output data are calculated on input data from sensors and other nodes and sent to target fog nodes which can calculate on the output data. In this article, we consider the mobile topic‐based publish/subscribe fog computing (MPSFC) model where mobile fog nodes communicate with one another by publishing and subscribing messages in wireless networks. Subscription topics of a fog node denote input data on which the node can calculate output data. Publication topics of a message show data carried by the message. In the topic‐based data transmission (TBDT) protocol proposed in our previous studies, while a fewer number of messages are transmitted, the delivery ratio of messages is smaller than the epidemic routing protocol. In this article, we newly propose a pair of time‐to‐live‐based data transmission (TTLBDT) and probability and topic‐based data transmission (PTBDT) protocols in order to increase the delivery ratio. If another node is found in the communication range, a fog node forwards messages to the node. Even if the node is not a target node, the node receives the message with some probability in the PTBDT protocol. In the evaluation, we show the delivery ratios in the TTLBDT and PTBDT protocols are larger than the TBDT protocol. In addition, we show the number of messages exchanged in the PTBDT protocol is smaller than the epidemic routing protocol.

Summary In this paper, we propose a P2PPSO (P2P [peer‐to‐peer] type of topic‐based Publish/Subscribe with Object concept) model, where each peer process (peer) exchanges objects by publishing and receiving event messages with no centralized coordinator. In addition, objects are characterized by topics. Suppose an event message ej published by a peer pj carries an object on some topics into a target peer pi. Here, information in the peer pj illegally flows to the peer pi if the target peer pi is not allowed to subscribe the topics. An illegal object is an object whose topics a target peer is not allowed to subscribe. Even if an event message is received by a target peer by checking topics, objects carried by the event message may be illegal at the target peer. In this paper, we propose a TOBS (topics‐of‐objects–based synchronization) protocol to prevent target peers from receiving illegal objects in the P2PPSO model. Here, even if an event message is received by a target peer, illegal objects in the event message are not delivered to the target peer. In the evaluation, we show how many event messages carry illegal objects in the TOBS protocol.

Processes on virtual machines can migrate from a host server to a guest server by migrating the virtual machines. In this paper, we propose a DMMV (Dynamic Migration of Multiple Virtual machines) algorithm where virtual machines migrate from a host server to a more energy-efficient guest server. Here, virtual machines are dynamically suspended and resumed depending on the number of processes. In addition, one or more than one virtual machine is selected to migrate from a host server to a guest server. The number of virtual machines to migrate is decided so as to minimize the electric energy consumption of the host and guest servers. In the evaluation, we show not only the total electric energy consumption and active time of servers but also the average execution time of processes can be reduced in the DMMV algorithm compared with other algorithms.

Server cluster systems equipped with virtual machines are widely used to realize fault-tolerant systems. Application processes are redundantly performed on multiple virtual machines to realize reliable distributed applications. However, a large amount of electric energy is consumed in a server cluster system since multiple replicas of each application process are redundantly performed on multiple virtual machines. In this paper, the redundant active time-based (RATB) algorithm is proposed to select multiple virtual machines for each application process so that the total electric energy of a server cluster to redundantly perform each application process can be reduced. In the RATB algorithm, it is not necessary for a load balancer to communicate with every virtual machine to estimate the electric energy of each server. Hence, the overhead of a load balancer to estimate the electric energy of each server and forward request processes can be reduced. We evaluate the RATB algorithm in terms of the total electric energy of a server cluster and the average response time of each process compared with the redundant energy consumption laxity based (RECLB) algorithm.

In the peer-to-peer (P2P) type of topic-based publish/subscribe (P2PPS) model, each peer process (peer) can publish and subscribe event messages which are characterized by topics with no centralized coordinator. A peer pj publishes an event message ej after receiving an event message ei, i.e. ei causally precedes ej. Here, the event message ej may carry information of the event message ei. If a peer pk receiving ej is not allowed to subscribe the topics of ei, the peer pk illegally obtains the information by receiving ej. In our previous studies, the SBS, TBS, and FS-H protocols are proposed to prevent illegal information flow. In addition, if a source peer pi publishes an event message e without giving related topics or with giving unrelated topics, a destination peer pj may misunderstand the meaning of the event message e. Here, malicious information flow occurs. In this paper, we newly propose a TBSM (topic-based synchronization to prevent malicious information flow) protocol. Here, event messages which may cause malicious information flow are banned. In the evaluation, we show the number of event messages banned in the TBSM protocol is larger than the TBS protocol since the number of event messages which are just malicious is larger than just illegal.

It is critical to reduce the electric energy consumption of servers in a cluster. In this paper, we discuss a migration approach to reducing the electric energy consumption of servers where a virtual machine with application processes migrates to a more energy-efficient server. In this paper, we consider a homogeneous cluster. We newly propose an ISEAM2H algorithm. Here, a virtual machine on a host server is selected to perform a process issued by an application and a guest server to which a virtual machine migrates is found so that the total electric energy consumption of servers can be minimized. In the evaluation, we show the total electric energy consumption and active time of the servers and the average execution time of processes can be reduced in the ISEAM2H algorithm.

intelligent, complex, reliable, and optimized cloud computing systems with Internet of Things (IoT) technologies among researchers, developers, and industrial experts.

Virtual machines are now widely used to support applications with virtual computation service in server clusters. Here, a virtual machine can migrate to a guest server while processes are being performed. In this paper, we discuss a virtual machine migration approach to reducing the electric energy consumption of servers. We newly propose two types of algorithms, energy-aware virtual machine selection (EVMS) algorithm to select a virtual machine to perform a process newly issued by a client and energy-aware virtual machine migration (EVMM) algorithm to migrate a virtual machine to a guest server. Here, the termination time of each virtual machine is estimated without considering each process. We evaluate the EVMS and EVMM algorithms and show the total electric energy consumption and active time of servers and the average execution time of processes can be reduced compared with other non-migration algorithms.

Server cluster systems are widely used to provide not only scalable but also fault-tolerant distributed application services. Each application process has to be reliably performed in presence of server fault to provide reliable application services. One way to reliably perform an application process is that replicas of each application process are performed on multiple servers. However, a larger electric energy is consumed in a server cluster since multiple replicas are performed on multiple servers. In our previous studies, the improved redundant delay time-based (IRDTB) algorithm is proposed to select multiple servers for redundantly and energy efficiently performing computation type application processes in a server cluster. In the IRDTB algorithm, a request process broadcasts to multiple servers and replicas for the request process starts almost at the same time on each server. Here, once a replica successfully terminates on one server, replicas being or to be performed on other servers are meaningless. In this paper, we newly propose an extended IRDTB (EIRDTB) algorithm to furthermore reduce the total energy consumption of a server cluster by forcing meaningless replicas to terminate and differentiating the starting time of each replica. In the EIRDTB algorithm, a request process is serially sent to each server every δ time unit. We evaluate the EIRDTB algorithm in terms of total energy consumption of a server cluster and the average response time of each process compared with the IRDTB algorithm. The evaluation results show the longer inter-request time δ, the smaller amount electric energy is consumed in the EIRDTB algorithm than the IRDTB algorithm.

In the peer-to-peer type of topic-based publish/subscribe (P2PPS) model, each peer (process) can publish and receive event messages with no centralized coordinator. A peer is allowed to publish and subscribe an event message with topics which are in the publication and subscription topics of the peer. Some information may flow from a peer to another peer if the peer publishes an event message to the other peer. We have to prevent illegal information flow to occur by publications and notifications of event messages. In our previous studies, the subscription-based synchronization (SBS), subscription initialization SBS (SI-SBS), topic-based synchronization (TBS), and SI-TBS protocols are proposed. In the SBS and SI-SBS protocols, it is checked whether or not illegal information flow to occur in terms of subscription and publication rights granted to each peer. However, even some legal notifications are banned while no illegal event message is notified. In the TBS and SI-TBS protocols, only topics which each peer manipulates are considered. In this paper, we evaluate the protocols in terms of the number of notifications banned.

In clusters of servers like cloud computing systems, computation resources like CPUs and storages are virtualized. Virtual machines are now widely used to support applications with virtual computation service on computation resources. Application processes are performed on virtual machines independently of which servers of which architectures are used. Furthermore, a virtual machine can migrate from a host server to a guest server while processes are being performed on the virtual machines. In this paper, we take advantage of the migration technologies of virtual machines to reduce the electric energy consumed by servers. We propose a modified simple virtual machine migration (MSVM) algorithm to migrate a virtual machine to another energy-efficient server in order to reduce the electric energy consumption. Here, the amount of computation to be performed by processes on a virtual machine is estimated only by using the number of the processes. We show the total electric energy consumption of the servers can be reduced in the MSVM algorithm compared with non-migration algorithms.

We have to reduce the electric energy consumed by servers in a cluster. In our previous studies, types of algorithms to select an energy-efficient server for a request process are proposed. Furthermore, schemes for energy-efficiently migrating a process and replicas of a process are discussed where a process and a replica migrate on a host server to a more energy-efficient guest server in our previous studies. However, it is not easy to realize the migration of processes on types of servers. Virtual machine (VM) technologies are now widely used to support applications with virtual resource service in could computing systems. Here, a virtual machine with application processes can migrate from a host server to another guest server. In this paper, we newly propose an energy-aware migration scheme of virtual machines (EAMV). Here, processes on a virtual machine can migrate to a server which consumes smaller electric energy and can be energy-efficiently performed in a cluster. We implement the EAMV scheme and evaluate the migration scheme in terms of energy consumption of servers and execution time of each process.

Scalable, high performance, and fault-tolerant distributed applications are realized with virtual machines in server cluster systems. Application processes are performed on virtual machines in each server. Processing load of virtual machines to perform application processes has to be balanced in a server cluster system to satisfy the application requirements like response time. On the other hand, a server cluster system consumes a large amount of electric energy since multiple servers consume electric energy to perform application processes. It is critical to discuss how to reduce the total electric energy consumption of a server cluster to perform application processes on virtual machines. In our previous studies, the transmission model and power consumption model of a server to perform communication processes on multiple virtual machines are proposed. In this paper, we newly propose the transmission energy consumption laxity based (TECLB) algorithm to allocate communication processes to virtual machines in a server cluster based on the proposed transmission model and power consumption model of a server so that the total energy consumption of a server cluster can be reduced. We evaluate the TECLB algorithm in terms of the total energy consumption of a server cluster and transmission time of each process compared with the basic round-robin (RR) algorithm. The evaluation results show the average total energy consumption of a server cluster is maximumly reduced to 9% of the RR algorithm.

It is critical to reduce the electric energy consumed in information systems, especially server clusters. In this paper, we extend the multi-level power consumption (MLPC) model and the multi-level computation (MLC) model to a server with multiple CPUs. In this paper, we newly propose a totally energy-aware (TEA) algorithm to select a server for a process in a cluster. Here, servers in a cluster are first classified into subclusters. Each subcluster is characterized in terms of the electric power and computation rate. One server is randomly selected in each subcluster. Then, one server is selected so that the expected electric energy is minimum in the selected servers. We evaluate the TEA algorithm and show not only the total electric energy consumption of the servers but also the average execution time of processes are reduced in the TEA algorithm compared with other algorithms.

Data in an object may flow into another object if transactions read and write data in the objects. A transaction illegally reads data in an object if the object includes data in other objects which are not allowed to be read. In our previous studies, the flexible read-write-abortion with role sensitivity (FRWA-R), object sensitivity (FRWA-O), and role safety (FRWA-RS) protocols are proposed to prevent illegal information flow. Here, a transaction aborts with some probability once illegally reading data in an object. The abortionprobability of a transaction depends on the sensitivities of roles which the transaction holds and objects in which the transaction illegally reads data. The safety of a role and object shows how many transactions which hold the role and illegally read data in the object commit or abort after illegally reading data in the object, respectively. Here, safety of a role which increases and decreases by constant values as a transaction holding the role aborts and commits, respectively. Based on the safety concept, we newly propose an influential abortion probability of a transaction where the abortion probability depends on not only roles held by the transaction but also the previous abortion probability. In this paper, we newly propose an FRWA with the influential abortion probability (FRWA-IAP) protocol. In the evaluation, we show a fewer number of transactions abort in the FRWA-IAP protocol than the RWA and FRWA-RS protocols while transactions are more efficiently performed than the WA protocol.

A distributed system is considered in an event-driven model where a state of a process transits on occurrence of an event. In this paper, we discuss a peer-to-peer topic-based publish/subscribe (P2PPS) model where each peer process (peer) can both subscribe and publish event messages. The subscriptions and publications are specified in terms of topics. Each event message e carries a vector e:TV = (tv1, , tvh) of topics t1, , th in a system. An event message e1 causally precedes an event message e2 with respect to a topic subset T iff not only e1 causally precedes e2 but also e1.tvj ≤ e2.tvj for every topic tj in an intersection T of the publications of e1 and e2 and the subscription Si of pi. A pair of event messages e1 and e2 are unnecessarily ordered if e1.TV ≤ e2.TV but e1 does not causally precede e2. In this paper, we newly propose a topic-based-causally delivering (TBC) protocol where the linear clock and physical clock are used with topic vectors to reduce the number of pairs of messages unnecessarily ordered. We evaluate the TBC protocol and show the number of pairs of unnecessarily ordered messages is reduced.

It is now critical to reduce electric energy consumed in a cluster of servers, especially scalable systems like cloud computing systems. In clusters, most application processes like web applications use not only CPU resources but also files and databases. In this paper, we consider storage processes which read and write data in files in addition to computation processes. We propose a PCS model (power consumption model for a storage server) which shows how much electric power a server consumes to perform storage and computation processes. We also propose a CS model (a computation model for storage server) which shows how long it is expected to take to perform storage processes and computation processes. By using the PCS and CS models, we propose a local energy-aware (LEA) algorithm to select a server for a request process in a cluster so that the total electric energy consumption of the servers can be reduced. We evaluate the LEA algorithm in terms of total electric energy consumption of the servers. We show the electric energy consumed by servers to perform computation and storage processes can be reduced in the LEA algorithm.

We have to reduce the electric energy consumed by servers in a cluster in order to realize eco-society. Types of algorithms for a request process to select an energy-efficient server in a cluster of servers are proposed in our previous studies. Furthermore, algorithms for energy-efficiently migrating a process on a host server to a more energy-efficient guest server is discussed. Virtual machines are now widely used to support applications with virtual computation service in cloud computing systems. Here, a virtual machine can migrate to a guest server, e.g. which is less loaded. By migrating a virtual machine, application processes performed on the virtual machine can also migrate from a host server to another guest server. In this paper, we newly propose an energy-aware migration algorithm of virtual machines (EAMV). Here, processes on a virtual machine can migrate to a guest server which consumes smaller electric energy and can be energy-efficiently performed in a cluster. We evaluate the EAMV algorithm compared with non-migration algorithms in terms of the total electric energy consumption and execution time of processes. We show the electric energy consumption and average execution time can be reduced in the EAMV algorithm.

Scalable and high performance server cluster systems like cloud computing systems are equipped with virtual machines to efficiently utilize server resources. On the other hand, a large amount of electric energy is consumed in a server cluster system since multiple servers consume electric energy to perform application processes on virtual machines. In order to design and implement an energy-aware server cluster system, the computation model of a virtual machine and power consumption model of a server have to be defined. In our previous studies, we proposed the power consumption model of a server and computation model of a virtual machine to perform application processes on virtual machines. However, the proposed power consumption model and computation model does not consider the change of the clock frequency of each core in a server. In this paper, we consider the change of the clock frequency of each core in a server. Then, we newly proposed the extended power consumption model of a server and computation model of a virtual machine to perform computation type application processes.

It is critical to reduce the electric energy consumed in information systems, especially server clusters. In this paper, we discuss an MLPCM (multi-level power consumption with multiple CPUs) model and an MLCM (multi-level computation with multiple CPUs) model of a server with multiple CPUs. In this paper, we newly propose a modified globally energy-aware (MEA) algorithm to select a server for a process in a cluster of m servers. In the MEA algorithm, a server where a process all is to be performed is selected with computation complexity O(m) if the total electric energy of the servers is minimum. We evaluate the MEA algorithm and show not only the total electric energy consumption of the servers but also the average execution time of processes are reduced in the MEA algorithm compared with other algorithms.

A publish/subscribe (PS) model is an event-driven model of a distributed system. In this paper, we consider a peer-to-peer (P2P) type of PS model where each peer (process) can publish and subscribe events. Here, a peer publishes an event message and then the event message is notified to a target peer which is interested in the event. Publications and subscriptions are specified in terms of topics as discussed in topic-based PS systems. In this paper, we newly discuss a topic-based access control (TBAC) model to prevent illegal information flow among peers in PS systems. Here, an access right is a pair "t, op" of a topic t and an operation op which is publish or subscribe. A peer is allowed to publish an event message with topics and subscribe topics only if the topics are granted to the peer. An event message e is notified to a peer pi if the publication of e and subscription of pi include some common topic. If a peer pi publishes an event message e2 after receiving an event message e1, the event message e2 may bring the event of e1, which the peer pi is not allowed to publish. Here, information in the peer pi illegally flow to another peer. We define the legal flow relation among the peers. Then, we newly propose a subscription-based synchronization (SBS) protocol to prevent illegal information flow. Here, a notification is banned if the notification may cause illegal information flow. We evaluate the SBS protocol in terms of number of notifications banned.

A distributed system is considered in an event-driven model composed of processes which are sequences of events. In this paper, we discuss a peer-to-peer (P2P) model of a topic-based publish/subscribe system (P2PPS model) where each peer process (peer) can both subscribe interesting topics and publish event messages. Since each peer can both publish and receive event messages, a peer may publish an event message e1 after receiving an event message e2. In the P2PPS system, event messages have to be notified to every common target peer in the causal order. In the TBC (topicbased causally delivery) protocol, event messages are causally delivered by taking advantage of physical time in addition to linear time and topic vector. Physical time is used to reduce pairs of event messages unnecessarily ordered. In this paper, we discuss the relation among the clock offset and delay time in the evaluation.

Application processes have to be efficiently and reliably performed on servers in a cluster. A process is replicated to increase the reliability and availability. However, the more number of replicas of a process are performed, the more reliable and available the system is but the more amount of electric energy is consumed. In this paper, we take a process migration (MG) approach to energy-efficiently and reliably performing an application process on servers in a cluster. A process performed on a current server is migrated to another server if the server is expected to consume a smaller amount of electric energy to perform the process than the current server while the deadline constraint of the process is satisfied. In order to reliably perform a process, multiple replicas of the process are performed on different servers. We consider synchronous and asynchronous ways to migrate process replicas to servers. In the synchronous migration, every replica moves to another server on a same computation state. In the asynchronous migration, each replica by itself makes a decision on when the replica is migrated to another server.

Various types of scalable and high performance computing systems are developed to realize distributed applications. In order to realize not only scalable but also high performance information systems, serverclusterc systems are widely used. In server cluster systems, the larger electric energy is consumed since application processes are performed on the larger number of servers. The improved delay time-based (IDTB)algorithm is discussed to select a server for each request process in a homogeneous server cluster so that the total energy consumption of the homogeneous server cluster to perform application processes can be reduced. In a homogeneous cluster, every server follows the same computation model and power consumption model. In heterogeneous cluster, servers follow different types of computation models and power consumption models. In this paper, we propose an extended IDTB (EIDTB) algorithm to reduce the total energy consumption and response time of each process in heterogeneous server cluster. We evaluate the EIDTB algorithm compared with the IDTB algorithm in a heterogeneous server cluster. In the evaluation, we show the average total energy consumption of a heterogeneous server cluster in the EIDTB algorithm can be reduced maximum 57% of the IDTB algorithm.

Information systems have to be secure in presence of illegal access. A transaction illegally reads an object if the transaction reads the object which includes data in other objects which is not allowed to be read. A transaction illegally writes an object after illegally reading some object. In addition, we consider suspicious object whose data is not allowed to flow to another object. A transaction suspiciously reads a suspicious object. A transaction impossibly writes an object after reading a suspicious object. Write-abortion (WA) and read-write abortion (RWA) protocols to prevent illegal information flow are already discussed. In the WA protocol, a transaction is aborted once issuing an illegal or impossible write. Reads are meaninglessly performed since the reads are undone due to the abortion of the transaction. In the RWA protocol, a transaction is aborted once issuing an illegal read or impossible write. Here, read operations to be performed after an illegal read are lost since a transaction is aborted just on issuing an illegal read. In this paper, we newly propose a flexible read-write abortion (FRWA) protocol to reduce the number of meaningless and lost reads. Here, a transaction is aborted with some probability if the transaction illegally reads an object. We evaluate the FRWA protocols compared with the WA and RWA protocols. We show the execution time of each transaction in the FRWA protocols is shorter than the WA and more number of reads can be performed in the RWA protocols.

In distributed systems, a group of multiple processes are cooperating with one another by exchanging messages in networks. A process is modeled to be a finite state machine. In this paper, we discuss a peer-to-peer (P2P) model of a publish/subscribe (P2PPS) system composed of peer processes (peers). Each peer can both subscribe a subscription and publish event messages with a publication. In this paper, subscriptions and publications are specified in terms of keywords. If a subscription of a subscriber peer and a publication of an event message include some common keywords, the subscriber peer is a target peer of the event message. The event message is notified to the target subscriber peer. A pair of event messages are related, which have a common target subscriber peer. Only a pair of related event messages are required to be delivered to common target subscriber peers in the causal order. We newly propose vectors of 〈V1, ..., Vm〉 of keywords k , ..., kmto causally order event messages. Each event message e carries the keyword vector e.V. An event message e1causally precedes an event message e2with respect to a subscription Siiff e1·Vh2·Vhfor every keyword khwhich is in the publications of the event messages e1and e2and the subscription Si. Only a pair of related messages are causally delivered to common subscriber peers.

Application processes have to be efficiently performed in presence of server faults in a cluster. Multiple replicas of a process are performed on multiple servers. However, the more number of replicas of a process are performed, the more reliable and available the process can be performed but the more amount of electric energy is consumed. In this paper, we take a process migration (MG) approach to energy efficiently and reliably performing multiple replicas of an application process on servers. A replica of a process performed on a current server st migrates to another server su if the server su is expected to consume a smaller amount of electric energy to perform the replica than the current server st while the deadline constraint of the process is satisfied. We consider synchronous and asynchronous ways for process replicas to migrate to servers. In the synchronous migration, every replica migrates to another server on a same computation state. In the asynchronous migration, each replica makes a decision on when the replica migrates to which server.

Scalable and fault-tolerant information systems like cloud systems are realized in server cluster systems. Server cluster systems are equipped with virtual machines to provide applications with scalable and fault-tolerant services. Scalable and fault-tolerant application services can be provided by balancing processing load among virtual machines to perform application processes. On the other hand, a large amount of electric energy is consumed in a server cluster system since multiple virtual machines are performed on multiple servers which consume electric energy to perform application processes. In order to design and implement an energy-aware server cluster system, the computation model and power consumption model of a server to perform application processes on multiple virtual machines have to be defined. In this paper, we first define the computation model of a virtual machine to perform application processes. We also define the power consumption model of a server to perform application processes on virtual machines.

The electric power consumed by servers has to be reduced in a cluster in order to realize eco society. We take a macro level approach to reducing the total electric energy consumption of servers to perform application processes in a server cluster. Servers are now equipped with multi-core CPUs. In this paper, we discuss a multi-level power consumption (MLPC) model of a server with a multi-core CPU. Here, the power consumption of a server depends on the number of active cores and active threads where at least one application process is performed. We also discuss a multi-level computation (MLC) model which gives the expected execution time of a process which is concurrently performed with other processes on a server with a multi-core CPU. Based on the MLPC model and the MLC model, we discuss an energy-aware (EA) algorithm to select a server for each process requested by a client in a cluster so as to reduce the total electric energy consumption while satisfying deadline requirements of the processes. We evaluate the EA algorithm and show not only the total energy consumption but also the average execution time of each process is reduced in the EA algorithm compared with the round-robin (RR) and random (RD) algorithms.

In information systems, objects have to be not only legally accessed in presence of illegal accesses but also data in a suspicious object is not allowed to flow to another object. A transaction illegally and suspiciously reads an object if the transaction reads the object which includes data in other objects which are not allowed to be read and the object which is suspicious, respectively. A transaction illegally and impossibly writes an object after illegally and suspiciously reading an object, respectively. The write-abortion (WA), read-write-abortion (RWA), and flexible RWA (FRWA) protocols to prevent illegal information flow are proposed in our previous papers. In the WA and RWA protocols, a transaction is aborted once issuing an illegal or impossible write and issuing an illegal read or impossible write, respectively. Reads are meaninglessly performed since the reads are later undone due to the abortion of the transaction in the WA protocol. Reads which can be performed after an illegal read are not performed, i.e. Lost since a transaction is aborted just once issuing an illegal read in the RWA protocol. In the FRWA protocol, a transaction is aborted with some probability once illegally reading an object in order to reduce the number of meaningless and lost reads. We newly introduce the sensitivity concept of an object in order to decide on the abortion probability. The more number of transactions are aborted after illegally reading an object, the more highly sensitive the object is. A transaction which illegally reads a more sensitive object is aborted with higher probability. In the evaluation, we show the execution time of each transaction in the FRWA protocol is shorter than WA and more number of reads can be performed than RWA.

Application processes are performed on servers in a cluster. Here, processes have to be not only reliably performed in presence of server faults but also the total electric power consumed by servers has to be reduced. The more number of replicas, the more reliably a process can be performed but the more amount of electric power is consumed by the servers. In this paper, we consider a mobile agent approach to realizing fault-tolerant energy-aware clusters. Here, a mobile agent is passively replicated on multiple servers. One primary replica of a mobile agent moves around servers and is performed on servers while the other secondary replicas are not performed. If the primary replica is faulty, one of the secondary replicas takes over the primary replica. In this paper, we discuss how to select servers for replicas to reliably and energy-efficiently perform mobile agents. We evaluate the energy-efficient passive replication scheme of a mobile agent in terms of average execution time of each mobile agent.

Application processes have to be not only fault-tolerantly but also energy-efficiently performed in presence of server faults in a cluster of servers. In our previous studies, the extended improved redundant power consumption laxity-based (EIRPCLB) algorithm is proposed where a process is actively replicated on multiple servers. While the cluster can be fault-tolerant, the larger amount of electric power is consumed than non-redundant execution. In order to reduce the total power consumption of servers, the total computation time of the replicas has to be reduced. We discuss an EIRPCLB algorithm for force termination and late restart of replicas. Here, as long as one replica successfully terminates, only a smaller part of every other replica is performed. In this paper, we evaluate the EIRPCLB algorithm in terms of total power consumption and the average response time in homogeneous and heterogeneous clusters. We make clear how the total power consumption of a cluster and response time of each process change according to the redundancy of each process and delay time between servers.

In information systems, data in an object may illegally flow into another object if a subject manipulates the objects. In this paper, we discuss information flow control models to prevent illegal information to occur in the role-based access control (RBAC) model. First, we define a legal information flow relation ri-> rj among roles ri and rj. It means, if a subject granted the role ri manipulates objects before another subject granted the role rj, no illegal information flow occur. We discuss safe systems where no illegal information flow occur even if operations from different subjects are performed in any order. Then, we discuss a role-based synchronization (RBS) protocol and an object-based synchronization (OBS) protocol to prevent illegal information flow in unsafe systems. Here, a transaction is aborted if the transaction reads an object and illegal information flow might occur. In the RBS protocol, the illegal information flow condition is specified in terms of roles while objects in the OBS protocol. We evaluate the RBS and OBS protocols in terms of number of transactions aborted.

Application processes have to be efficiently and reliably performed on servers in a cluster. Multiple replicas of a process can be performed on multiple servers in order to make the process fault-tolerant. The more number of replicas are performed on multiple servers, the more reliably a process can be performed but the more amount of electric power is consumed by the servers. In this paper, we consider a mobile agent approach to fault-tolerantly, energy-efficiently performing application processes in a cluster. Here, a process is first initialized on a server named home server. A process is thus performed on a server st and moves to another server su if the server su spends smaller power to perform the process than st. A process takes checkpoints and sends the checkpoints to the home server. If a process is faulty, the home server recreates the process an operational server and the process is restarted on a state saved at a checkpoint most recently taken on the home server.

The improved power consumption laxity-based (IPCLB) algorithm is discussed to select one of servers so that the total power consumption of a cluster can be reduced. However, a load balancer has to collect a state of every current process on servers of a cluster to calculate the estimated power consumption of each server. In addition, it is difficult to precisely estimate the power consumption of each server since the state of each process on the server is changed during the estimation. Especially, a process might terminate before the termination time is estimated if the computation time of the process is shorter than the communication delay time between the load balancer and the server. In this paper, we assume the computation time of each process is shorter than the communication delay time. Then, we propose a delay time-based (DTB) algorithm to select a server for each request process so that the total power consumption of a cluster to perform processes on the server can be reduced. In the DTB algorithm, it is not necessary to collect a state of every process on each server to estimate the power consumption laxity. In addition, the minimum computation time of a process is not required to be a priori defined in the DTB algorithm.

In information systems, data in an object may illegally flow into another object if transactions manipulate the objects. In our previous studies, the role-based synchronization (RBS) and object-based synchronization (OBS) protocols are discussed to present illegal information flow in the role-based access control (RBAC) model. Illegal read operations mean read operations which might imply illegal information flow. Transactions which issue illegal read operations are aborted. In this paper, we consider a unique object whose data is not allowed to flow to another object. An illegal write is defined to be a write operation which is issued after reading a unique object or illegally reading an object in a transaction. Each transaction reads any object but is aborted if the transaction illegally writes an object. Synchronization protocol WA-RBS and WA-OBS are discussed based on abortion of transaction issuing illegal write operations.

In distributed systems, a group of multiple processes are cooperating with each other through networks. A process is modeled to be a sequence of events in a finite state machine model. Here, it is significant to decide on which event occurs before another event for a pair of events in a system. In the famous happened-before relation among events, events are ordered on the basis of send-receive semantics of messages. Only events which are meaningful in distributed applications should rather be causally ordered, i.e. events which occur in some situation. We consider publish/subscribe (PS) systems which are event-driven systems and are used in various distributed applications. The causal order of events is important to be discussed in a situation where the events meaningfully occur. We consider a peer-to-peer (P2P) model of a publish/subscribe system (P2PPS) where every peer process (peer) can publish events, subscribe events, and is notified of events. There are subscribing, publishing, and notifying events in P2PPS systems. In this paper, we discuss a situation-aware causal dependency of application events in a P2PPS system.

In information systems, applications are required to be realized in the digital ecosystems, high performance systems, and scalable systems. In these applications, a client first selects a server in a cluster of servers and issues a request to the server. The request is performed as a process in the server. In this paper, we consider a computation process which mainly consumes the central processing unit (CPU) resources. Cooling devices, such as fans, consume the electric power in addition to the CPU of a server. The rotation speed of a fan is revved up in servers in order to decrease the temperature of a server. Thus, the total power consumption of a server depends on not only computational devices such as a CPU, but also cooling devices. We discuss an extended simple power consumption (ESPC) model of a server with a multicore CPU and cooling devices to perform computation type processes. It is critical to discuss how to select a server for each request issued by a client in order to not only achieve performance objectives, but also reduce the total power consumption of a system based on the ESPC model. An improved power consumption laxity-based (IPCLB) algorithm for selecting a server is proposed in this paper, where the minimum power to be consumed is used to perform the process. We evaluate the ESPC model and IPCLB algorithm in terms of power consumption and elapse time.

It is critical to discuss how to realize not only energy-aware but also robust clusters of servers. A client usually issues a request to one server in a cluster and the server sends a reply to the client. Once the server stops by fault, the client does not receive a reply of the request and might be suspended to wait for a reply. Hence, each request is redundantly performed on multiple servers to be tolerant of server faults. In our previous studies, multiple servers are selected to redundantly and energy-efficiently perform a request process in the redundant power consumption laxity-based (RPCLB) algorithm. Here, since each application process is redundantly performed on more than one server, the larger amount of electric power is consumed. In this paper, we newly propose the improved RPCLB (IRPCLB) algorithm where once a process successfully terminates on one server, meaningless redundant processes are not performed on the other servers. We show the total power consumption of servers is reduced in the IRPCLB algorithm while the execution time of processes is almost the same as the RPCLB and round-robin (RR) algorithms.

Electric power consumed by servers has to be reduced in order to realize green societies. We consider computation ({\it{CP}}) and storage ({\it{ST}}) types of application processes performed on servers in this paper, where CPU and storage drives are mainly used, respectively. In the storage and computation based power consumption (SCBPC) model proposed by the authors, the power consumption rate of a server depends on what types of processes are performed. In the storage and computation based processing (SCBP) model, the execution time of an ST process depends on the number of concurrent CP and ST processes but the execution time of a CP process depends on only CP processes. In our previous studies, the energy-aware (EA) algorithm is discussed to select a server in a cluster of servers for each request so that the total power consumption of the servers can be reduced. However, a server consumes electric power even if no process is performed. In this paper, we discuss a dynamic EA (DEA) algorithm where the membership of a cluster is dynamically changed so that only active servers where at least one process is performed are included in the cluster. A server for each request is selected in a dynamic cluster so that the total power consumption of servers in the cluster can be reduced. We evaluate the DEA algorithm in terms of the total power consumption and average execution time.

We discuss how to efficiently and consistently manipulate multiple replicas of a multimedia object. Multimedia replicas are characterized in terms of not only data structure but also quality of service (QoS). Multimedia replicas are written in enriching and impoverishing types of write operations where some data is added and removed in a replica, respectively. Computation resources are spent to materialize, i.e. physically update multimedia replicas by encoding and decoding multimedia data. We discuss an extended multimedia quorum-based (EMQB) protocol to reduce the processing overhead of each replica. Here, a replica is not materialized in an impoverishing type of write operation. Even in an enriching type of write operation, only some number, not necessarily all of replicas in a write quorum are materialized. Hence, a newest, materialized replica may not be in a read quorum. Even if one newest replica is surely found in a read quorum, no newest replica might be materialized and a transaction cannot read any replica. In the EMQB protocol, replicas are randomly selected to be in a read quorum until a materialized replica is found. The EMQB protocol is evaluated in terms of the total processing overhead of replicas compared with the quorum-based (QB) protocols. We show that the processing overhead of each replica can be reduced in the EMQB protocol compared with the QB and MQB protocols.

Replicas of a multimedia object are characterized in terms of not only data structure but also quality of service (QoS). Multimedia replicas are updated by two types of write operations, i.e. enriching and impoverishing types, where some data is added and removed in a replica, respectively. Computation resources are spent to materialize, i.e. physically update multimedia replicas by encoding and decoding multimedia data. It is critical to reduce the processing overhead of each replica to materialize. In our protocol named an extended multimedia quorum-based (EMQB) protocol, a replica in a write quorum is not materialized each time a write operation is issued. Only in an enriching type of write operation, some number, not necessarily all of replicas in a write quorum are materialized. In the other replicas, write operations are just logged but not performed. The replica can be materialized by performing operations in the log if every operation is an impoverishing type because just some data in a physical state of the replica is removed. However, the replica cannot be materialized if some operation is an enriching type in the log. Hence, a newest, materialized replica may not be in a read quorum. In the EMQB protocol, replicas are randomly selected to be in a read quorum until a materialized replica is ound. We show that the processing overhead of each replica can be reduced in the EMQB protocol compared with the QB and MQB protocols in terms of the total processing overhead of replicas.

In peer-to-peer (P2P) overlay networks, a group of n (≥2) peer processes have to cooperate with each other. P2P systems are in nature scalable and distributed, Each peer sends messages to every peer and receives messages from every peer in a group. In group communications, each message sent by a peer is required to be causally delivered to every peer. Types of logical clocks are used to causally order messages, linear and vector clocks. The linear clock can be used in a scalable group since message length is O(1). However, some pair of messages not to be causally ordered are unnecessarily ordered. On the other hand, no messages are unnecessarily ordered in the vector clock. However, the vector clock cannot be adopted in a scalable group due to the message length O(n). Recently, processes on each computer can read more process is obtain more process more time by reading a physical clock is synchronized with time servers. In this paper, we discuss how to causally deliver messages in a scalable group by taking advantage of linear and physical clocks. Here, the number of pairs of messages unnecessarily ordered in the linear clock can be also reduced.

In a server cluster, one server is usually selected to perform a request process from a client. Once the server stops by fault, the client is suspended to wait for a reply. Even if the request is performed on another server on detection of fault of the server, some QoS requirement like response time may not be satisfied. Hence, each request is redundantly performed on multiple servers to be tolerant of server faults. Here, more number of servers a request process is redundantly performed, the more reliable but the more amount of electric power is consumed. Thus, it is critical to discuss how to realize a reliable and energy-aware server cluster in presence of server faults. The redundant power consumption laxity-based (RPCLB) algorithm is proposed to redundantly and energy-efficiently perform a request process in our previous studies. In this paper, we newly discuss an improved RPCLB (IRPCLB) algorithm where once a process successfully terminates on one server, meaningless redundant processes are forced to terminate on the other servers. We show the total power consumption of servers and response time of each process can be reduced in the IRPCLB algorithm than the RPCLB and round-robin (RR) algorithms in both heterogeneous and homogeneous clusters.

Electric power consumed by servers has to be reduced in order to realize green societies. A server has to be selected in a cluster of servers so that the total power consumption can be reduced. We consider computation (CP) and storage (ST) types of application processes performed on servers in this paper, where CPU and storage drives are mainly used, respectively. In our previous studies, the energy-aware (EA) algorithm is discussed to select a server in a cluster of servers for each request so that the total power consumption of the servers can be reduced. However, an idle server consumes electric power even if no process is performed. In this paper, we discuss a dynamic energy-aware, heterogeneous (DEA-H) cluster. Here, only servers required to perform request processes but no idle servers are included in each cluster. In a DEA-H cluster, servers are selected in a server pool if the traffic increases and idle servers leave for the server pool if the traffic decreases. A server for each request is selected in a heterogeneous DEA cluster so that the total power consumption of servers can be reduced. We evaluate the DEA algorithm on a heterogeneous cluster (DEA-H) in terms of the total power consumption and average execution time.

Processes requested by clients are performed on servers in a cluster of servers. Here, not only QoS (quality of service) requirements like response time are satisfied but also the total electric power consumed by servers to perform processes has to be reduced. Furthermore, each process has to be reliably performed in presence of server faults. In this paper, we consider a mobile agent where a process, i.e. mobile agent, is reliably performed by moving around servers in presence of server faults. A replica of a mobile agent is first performed on one primary server while the other replicas of the agent are sent to other secondary servers. While performing the primary agent replica, the server takes checkpoints of the primary agent replica and sends the checkpoints to the secondary agent replicas. If the primary replica is faulty due to the fault of the primary server, one of the secondary replicas takes over the primary replica. The more number of agent replicas are performed, the more reliably the agent can be performed but the more amount of electric power is consumed by the servers. Hence, it is critical to discuss how to reliably and energy-efficiently perform mobile agents on multiple servers. In this paper, we discuss how to reduce the total electric power consumption in a cluster where each agent is passively replicated on multiple servers. If the primary server is faulty, one of the secondary servers gets primary. We evaluate the energy-efficient passive replication (EPR) scheme of a mobile agent in terms of total power consumption and average execution time and response time of each mobile agent in presence of server fault.

The power consumption of information systems, especially servers has to be reduced to realize green eco-society. A client first selects a server in a collection of possible servers and issues a request to the server. The request is performed as a process in the server. In this paper, we consider a computation type of application where CPU resource is mainly used in a server. It is critical to discuss how to select a server for each request from clients in order to reduce the total power consumption. The power consumption laxity based (PCLB)algorithm for selecting a server is discussed in this paper, where a server to consume the minimum power to perform a process is selected. In a server, cooling devices like CPU fans consume the electronic power in addition to CPU and other devices. In our previous studies, we assume the cooling devices consume the constant power and propose the simple power consumption model. In the simple power consumption model, if at least one application process is performed on a server, the electric power is maximally consumed on the server. Otherwise, the electric power consumption is minimum. The rotation speed of a fan is controlled in servers based on the temperature of CPU and in the server's case. Thus, the total power consumption of a server depends on not only the power consumption of computational devices like CPU but also the power consumption of cooling devices like fans. In this paper, we extend the simple power consumption model to take into account the power consumption of cooling devices. We improve the PCLB algorithm based on the extended simple power consumption model.

The electric power consumption of servers has to be reduced in information systems in order to realize green societies. In information systems, servers mainly consume the electric power to process regrets from clients. We have to reduce the electric power consumed by servers. There are computation (CP), communication (CM), and storage (ST) types of applications to be performed on servers. In CP and CM applications, CPU and communication resources are mainly consumed, respectively. In this paper, we consider ST applications where storage drives are manipulated on a server in addition to CP applications. First, we measure the power consumption of a server to perform CP and ST types of application processes, C, R, and W processes which just compute, read files, and write files, respectively. Then, we discuss a power consumption model of a server by abstracting most essential parameters dominating the power consumption of the server from the experimental results. Here, the power consumption rate of a server is max um if at least one process is performed. However, the maximum power consumption rate of a server depends on types of processes concurrently performed on the server. In addition, we obtain an access rate model to show how long it takes to perform each type of process concurrently with other processes. By using both the power consumption model and the access rate model, we discuss algorithms from selecting a server so that not only the execution time but also the power consumption can be reduced.

Multimedia objects are distributed to peers through downloading and caching in peer-to-peer (P2P) overlay networks. In the quorum-based (QB) protocols, every replica of an object is not up-to-date and the replicas are totally ordered in the version counters. On the other hand, multimedia objects are characterized in terms of not only data structure but also quality of service (QoS) like number of colours. Replicas of a multimedia object are partially ordered in terms of the newness precedent relation on not only data structure but also QoS. An object is an encapsulation of data structure and abstract operations for manipulating the data structure. Operations are classified into write (W), read (R), and read-write (RW) types. An object is changed though two types of W/RW operations, enriching and impoverishing types. Some data is added to a replica in an enriching operation. On the other hand, some data in a replica is removed in an impoverishing operation. In order to reduce the overhead to write every replica in a quorum, we take an approach that every replica is not updated. If a transaction issues an enriching operation, every replica in the quorum is updated in the same way as the QB protocol. On the other hand, if an impoverishing type of write operation is issued, every replica is not updated in the quorum. Impoverishing operations are logged in replicas while enriching operations are performed on every replica. Then, impoverishing operations in a log are later performed on a replica.

In peer-to-peer (P2P) overlay networks, a group of multiple peers have to cooperate with each other. P2P systems are in nature scalable distributed systems, where there is no centralized coordinator. It is difficult, maybe impossible for each peer to communicate with every other peer in P2P overlay networks. An acquaintance peer of a peer is another peer with which the peer can directly communicate. Each peer has to obtain access and location information on resources like databases through communicating with acquaintance peers. It is critical to discuss how each peer can trust an acquaintance peer in P2P networks since acquaintance peers may have obsolete information and may be faulty. In this paper, we discuss subjective and objective types of trustworthiness of a peer on an acquaintance peer. A peer obtains the subjective trustworthiness on an acquaintance peer through directly communicating with the acquaintance peer. On the other hand, a peer obtains the objective trustworthiness on a target acquaintance peer through collecting subjective trustworthiness of other peers on the target acquaintance peer. In this paper, the trustworthiness is given based on the Fuzzy logics. There are cases the subjective and objective types of trustworthiness on an acquaintance peer are different. That is, other peers have different trustworthiness opinions on the target acquaintance peer. A peer decides on which type of trustworthiness to be taken based on the confidence. The confidence of a peer shows how much the peer is confident of its own trustworthiness opinion, i.e. subjective trustworthiness on the acquaintance peer. If a peer is confident of the trustworthiness opinion, i.e. the confidence is larger, the peer takes the subjective trustworthiness on the target acquaintance peer. Otherwise, the peer takes the objective trustworthiness.

In information systems, a client first selects a server in a cluster of servers and issues a request to the server. The request is performed as a process in the server. In this paper, we consider a computation process which mainly consumes CPU resource. Cooling devices like fans consume the electric power in addition to CPU and other devices in a server. The rotation speed of a fan is changed in servers according to the temperature of a server. Thus, the total power consumption of a server depends on not only computational devices like CPU but also cooling devices. The extended simple power consumption (ESPC) model of a server is proposed to perform processes. In this paper, the ESPC model is improved to consider a server with a multi-core CPU. It is critical to discuss how to select a server for each request from clients in order to not only achieve performance objectives but also reduce the total power consumption of a system based on the ESPC model. The improved power consumption laxity based (IPCLB) algorithm for selecting a server is proposed in this paper, which consumes the minimum power to perform the process. We evaluate the ESPC model and the IPCLB algorithm in terms of the power consumption and elapse time in this paper.

The electric power consumption of servers has to be reduced in information systems in addition to increasing the performance and fault-tolerancy in order to realize green societies. In information systems, servers mainly consume the electric power to perform processes requested by clients. We consider computation (CP) and storage (ST) types of application processes to be performed on servers. In CP processes, CPU resource is mainly consumed. In stprocesses, storage drives are manipulated in a server. First, we present the storage-based power consumption (SBPC) model of a server st to perform stand CP processes. The SBPC model is a four-state model, where the power consumption rate of the server is one of four rates minEt, maxSTt, maxCt, and maxSTCt. Here, the power consumption rate Et(τ) of a server st of time τ is minimum minEtif no process is performed. Et(τ) the power consumption rate is maxSTtif only and at least one ST process is performed. Et(τ) is maxCtif only and at least one CP process is performed on the server st. If both ST and CP processes are performed, Et(τ) is maxSTCt. Here, minEtttt. According to the SBPC model, it is better not to perform both ST and CP processes on a same server because the maximum power is consumed in the server. By using the SBPC model, we discuss an algorithm for selecting a server so that not only the execution time but also the power consumption can be reduced. In the algorithm, ST and CP processes are tried to be performed on different servers.

Replicas of multimedia objects are distributed to peers through downloading and caching in peer-to-peer(P2P) overlay networks. Every replica has to be consistent. In quorum-based (QB) protocols, every replica may not be up-to-date and the up-to-date replicas can be found in the version counters. On the other hand, multimedia objects are characterized in terms of not only data structure parameter but also quality of service (QoS) parameters like frame rates. Replicas of a multimedia object are partially ordered in terms of the newness precedent relation on data structure and QoS. Each parameter of a replica is manipulated in read and write operations. In addition, there is a read operation to read a whole replica. A parameter of an object is changed through a write operation. There are enriching and impoverishing types of write operations. Some data is added to a replica in an enriching operation. On the other hand, some data in a replica is removed in an impoverishing operation. In order to reduce the overhead to write every replica in a quorum, we take an approach that every replica is not updated. If a transaction issues an enriching operation, every replica in the quorum is updated in the same way as the QB protocol. On the other hand, if an impoverishing write operation is issued, every replica is not updated in the quorum. Impoverishing operations are just recorded in replicas. Then, impoverishing operations are later performed on a replica. The MQB protocol is evaluated in terms of the processing overhead in replicas.0

We discuss how to efficiently make replicas of a multimedia object consistent. Multimedia objects are characterized in terms of not only data structure but also quality of service (QoS). There are two types of write operations: enriching and impoverishing types. Some data is added to a replica in an enriching type while removed in an impoverishing type. Computation resources are spent to materialize, i.e. physically update replicas of a multimedia object. We discuss an extended multimedia quorum-based (EMQB) protocol to reduce the processing overhead of each replica. Here, a replica is not materialized in an impoverishing type of write operation. Even in an enriching type of write operation, only some number, not necessarily all, of replicas in a write quorum are updated. Hence, a newest materialized replica may not be found in a read quorum. Replicas are randomly selected to be in a read quorum until an updated replica is found in the quorum. The EMQB protocol is evaluated in terms of the total processing overhead of replicas. We show that the processing overhead of each replica can be reduced in the EMQB protocol.

Nowadays information systems are being shifted to distributed architectures, i.e. Grid and P2P models to obtain the benefits like scalability, autonomy, and fault-tolerance. We consider the peer-to-peer (P2P) model as a fully distributed, scalable system, which is composed of peer processes (peers). A group of multiple peers cooperate with each other. Here, peers have to efficiently and flexibly deliver messages to every peer of the group in P2P overlay networks. In order to efficiently and reliably broadcast messages in a scalable group, we take advantage of the multipoint relaying (MPR) mechanism. Here, each peer sends messages to only a subset of its acquaintances. However, if a peer to forward messages to other peers fails to forward the messages, the peers cannot receive the messages. In this paper, we newly discuss a trustworthiness-based broadcast (TBB) algorithm where only trustworthy peers forward messages. Even if untrustworthy peers do not forward messages, messages can be delivered to every peer. Here, the transmission fault implied by faults of untrustworthy peers can be reduced. We evaluate the TBB algorithm in terms of the number of messages transmitted.

In order to realize digital ecosystems, the total electric power consumption of computers and networks have to be reduced. In distributed systems, clients issue requests to servers and then servers send replies to clients. Here, application processes are composed of a pair of modules which mainly consume computation and communication resources, respectively. Based on the measurement of the power consumption of a server, we newly discuss a power consumption model of a server for applications. We propose an EPCLB (extended power consumption laxity-based) algorithm to select one of servers for applications so that the total power consumption of servers can be reduced. We evaluate the EPCLB algorithm in terms of execution time and power consumption compared with the traditional round-robin algorithm.

A group of n (>;1) peers are required to cooperate with each other in distributed applications. In this paper, we consider a distributed group of multiple peers in P2P overlay networks. A P2P group is distributed, i.e. no centralized controller and is scalable and heterogeneous since various types and huge number of computers are interconnected in types of networks. In group communication, messages have to be causally delivered to every peer. In order to realize a scalable group, messages are ordered by linear time (LT) and physical time (PT) since message length is O(1). In order to use the physical time, each peer has to hold information on the accuracy of physical clock of each peer and minimum delay time among every pair of peers. Hence, the size of the information is O (n2). In this paper, we newly discuss a multi-layered hierarchical model to reduce the size of group information. We evaluate the hierarchical model in terms of the group information size and delay time compared with a flat group model.

Information systems are being shifted to scalable architectures like Cloud and peer-to-peer (P2P) models. In this paper, we consider the P2P model as a fully distributed, scalable system different from centralized coordination systems in Cloud and Grid systems. A P2P system is composed of peer processes (peers). Here, applications are realized by activities and cooperations among multiple peers. In P2P systems, since there is no centralized coordination, each peer has to obtain information about others by itself. In the group activities, each group member has to be trustworthy so that malicious action of an individual peer cannot effect overall outcome of the whole group. It is important to consider the trustworthiness of each group member as a base of an agreement procedure in the distributed environment. The goal of the group and the way to archive the goal are decided by the group members. During the agreement procedure, opinion of each group member has to be considered. Malicious and unexpected behaviors of member peers can negatively effect the output of a group. Hence, it is significant to discuss how to compose a group only by including more trustworthy peers. In this paper, by taking advantage of the trustworthiness concept of each peer, we propose a novel approach to composing a trustworthy group in the distributed agreement protocols.

In information systems, it is critical to reduce the total electrical power consumption of computers and networks in order to realize the digital ecosystems and the green IT technologies. In the extended power consumption laxity-based (EPCLB) algorithm, a server is selected in a set of servers so as to not only satisfy deadline constraint but also reduce the total power consumption of servers in general types of applications. However, each time a load balancer receives a new request, the load balancer has to collect status of each server and calculate the estimated power consumption to perform the request. The load balancer spends large computation and communication overhead to estimate the power consumption if the number of clients is increased. In addition, since the status of each server might be changed during the estimation process, it is difficult to correctly estimate the power consumption. In this paper, we newly propose a CTRB (computation and transmission rate based) algorithm to select a server in a set of possible servers so that the total power consumption of servers and the overhead of a load balancer can be reduced. We evaluate the CTRB algorithm in terms of the power consumption of servers and the overhead of a load balancer compared with the EPCLB and traditional round-robin (RR) algorithms.

In order to realize digital ecosystems, the total electric power consumption of servers to perform applications has to be reduced in information systems. Applications are classified into computation, communication, and storage based types. In the computation and communication based applications, CPU and communication resources of computers are mainly consumed, respectively. In the storage-based applications, the electric power is consumed to access to storage drives like hard disk drives (HDDs) and solid state drives (SSDs). In this paper, we measure how much electric power of a server computer is consumed to read and write data in types of storage drives. Then, we discuss a power consumption model of a computer to perform a storage-based applications based on the measurement of the power consumption of storage drives. Here, the power consumption rate of a server is maximum if at least one process is performed. Otherwise, the power consumption rate is minimum. This is a simple power consumption model.

Information systems are getting scalable like cloud computing systems and peer-to-peer (P2P) overlay networks. In information systems, resource objects like databases are distributed and replicated in order to increase the performance, reliability, and availability. In this paper, we consider multimedia objects like movies which are characterized in terms of not only data structure like part of relations of objects but also quality of service (QoS) like frame rate. A pair of replicas may have the same content but different level of QoS. Thus, kinds of replicas of a multimedia object are distributed. Replicas of an object are partially ordered in terms of newness of contents and QoS parameters, while replicas are just totally ordered in traditional quorum theories. We classify operations on an object to three types I, O, IO types of operations with respect to whether or not data and QoS are read and written in the object. We define a quorum for each type of operation. We discuss how to synchronize multiple replicas on the basis of the quorum concept in scalable distributed systems.

A group of peers are required to cooperate with each other in distributed applications on P2P overlay networks. In order to realize a scalable P2P group, messages are causally ordered by taking advantage of linear time (LT) and physical time (PT) since message length is O(1) in this paper. In this paper, we newly discuss a multi-layered model to realize a scalable group. A group is hierarchically decomposed into subgroups and every peer is in a leaf peer. We present how to order messages by using PT and LT in a multi-layered group. We evaluate the multi-layered group protocol in terms of the group information size and delay time among peers compared with a flat group.

Multimedia objects are distributed to peers through downloading and caching in peer-to-peer (P2P) overlay networks. Multimedia objects are characterized in terms of not only data structure but also quality of service (QoS). Replicas of a multimedia object are partially ordered in terms of the newness precedent relation of data structure and QoS. In the multimedia quorum-based (MQB) protocol, multiple replicas of a multimedia object are synchronized based on the newness precedent relation. An object is an encapsulation of data and abstract operations for manipulating the data. The data of an object is changed in a write type of operation. Write operations are classified into enriching and impoverishing types of operations. In an enriching operation, some data is added to a replica like an operation to increase the number of colours. On the other hand, in an impoverishing operation, some data in a replica is removed like operation to degrade QoS. If an enriching operation is issued to replicas in a quorum, every replica in the quorum is updated like the traditional quorum-based (QB) protocols. On the other hand, if an impoverishing operation is issued to a quorum, every replica in the quorum is not updated where the operations are logged to reduce communication and computation overheads. In this paper, only impoverishing operations are logged while enriching operations are performed on every replica in the quorum. Then, only impoverishing operations to be performed in a log are later performed on a replica and logged in a replica. In addition the size of the log is reduced by removing operation not to be performed.

Abstract In peer‐to‐peer (P2P) applications, messages have to be causally delivered to every peer of a group ofn(>1) peers. Owing to the message overheadO(n), the vector clock cannot be used to causally deliver messages in a scalable group. On the other hand, since the message length isO(1) in a linear time (LT) and physical time (PT) protocols, which use the linear clock and physical clock, respectively, the protocols can be adopted to scalable groups. However, some pairs of messages are unnecessarily ordered, i.e. even if a pair of the messages is ordered in the protocols, the messages may not be causally ordered. The more the number of messages unnecessarily ordered, the longer it takes to deliver them. In this paper, we propose a hybrid time (HT) group communication protocol to reduce the number of messages unnecessarily ordered by taking advantage of the LT and PT. We evaluate the HT protocol compared with the PT and LT protocols in terms of the number of unnecessarily ordered messages. We show that the number of unnecessarily ordered messages can be reduced in the HT protocol. Copyright © 2010 John Wiley & Sons, Ltd.

Nowadays more and more information systems are being shifted to distributed architectures because of the benefits like scalability, autonomy, and faulty-tolerance implied from the essence of the distributed systems. Here, every process is peer and cooperates with other peers to achieve common goal. In order to do that, peers have to efficiently and flexibly make an agreement on one common value which satisfies an agreement condition. In this paper, we consider a distributed group of multiple peers with no centralized coordination. We introduce a novel approach to efficiently making an agreement where each peer sends a package of multiple possible values to the other peers at each ongoing round. By exchanging multiple possible values at once, we can significantly reduce the total number of messages. The time and network resources are mostly spent in the value exchange phase. If we can reduce the time and number of messages to exchange values among peers, we can improve the efficiency of the agreement protocol. In order to efficiently exchange value packages among peers, we take advantage of the multipoint relaying mechanism to reduce the number of duplicate re-transmissions. Although we can significantly reduce the re-transmitted values, we have to realize the fault-tolerancey of the system. In addition to improving the reliability of the multipoint relaying mechanism, we newly introduce the trustworthiness among peers. By taking into account the trustworthiness of the peer, each peer broadcasts values through the trusted neighbors to the other peers. Here, the transmission fault which causes by untrusted, unreliable peers can be prevented.

It is now critical to reduce the consumption of natural resources, especially petroleum. Even in information systems, we have to reduce the total electrical power consumption. We classify network applications to two types of applications, transaction and communication based ones. In this paper, we consider communication based applications like the file transfer protocol (FTP). A computer named server consumes the electric power to transfer a file to a client depending on the transmission rate. We discuss a model for power consumption of a data transfer application which depends on the total transmission rate and number of clients to which the server concurrently transmits files. A client has to find a server in a set of servers, each of which holds a file so that the power consumption of the server is reduced. We discuss a pair of algorithms PCB (power consumption-based) and TRB (transmission rate-based) to find a server which transmits a file to a client. In the evaluation, we show the total power consumption can be reduced by the algorithms compared with the traditional round-robin algorithm.

In peer-to-peer (P2P) applications, a group of multiple peer processes (peers) are required to cooperate with each other in a distributed manner. Messages sent by peers have to be causally delivered to every peer in a group. Due to the message overhead O(n) for the number n of peers, the vector clock cannot be used to causally deliver messages in a scalable group. On the other hand, the linear clock implies the message length O(1), but some pair of messages are unnecessarily ordered. Recently, more accurate physical clocks can be used in cooperation with the GPS time server. In this paper, we reduce the number of messages unnecessarily ordered by taking advantage of the linear time and physical time. Even if each physical clock is synchronized with a time server, every physical clock does not show the same accurate time. Let λishow the accuracy of the physical clock of a peer piand dijshow the minimum delay time between piand pj. In this paper, we consider a homogeneous broadcast group where λi= λ and dij= d for every pair of peers piand pjand each message is sent to every peer. In the evaluation, we show the number of messages unnecessarily ordered can be reduced in the protocol compared with the linear clock and the physical clock for δ ≤2λ.

Nowadays more and more information systems are being shifted to distributed architectures because of the benefits like scalability, autonomy, and faulty-tolerance implied from the essence of this distributed systems. Here, every process is peer and cooperates with other peers to achieve some objectives. Here, peers have to efficiently and flexibly make an agreement on one value which satisfies an agreement condition. We consider a distributed group of multiple peers with no centralized coordination. In order to efficiently make an agreement, we introduce a novel approach that each peer sends a package of multiple possible values to the other peers at each ongoing round. During the agreement procedure in traditional ways, every peer sends one value to the other peers at each round. It takes time for peers to exchange multiple values with each other and the peers might not be able to make an agreement depending on the order of values shown by the peers. By exchanging packages of values at one round, we can not only increase the possibility that every peer makes an agreement but also reduce the time to reach agreement and the number of messages exchanged among peers. In addition, peers have to efficiently broadcast packages of values to every peer in a group. In this paper, we newly introduce a broadcast message flooding algorithm by taking into account the multipoint relaying (MRPs) mechanism to increase the efficiency of the protocol.

It is now critical to reduce the consumption of natural resources, especially petroleum to resolve air pollutions. Even in information systems, we have to reduce the total electrical power consumption. A cloud computing system is composed of a huge number of server computers like google file systems. There are many discussions on how to reduce the total power consumption of servers, e. g. by turning off servers which are not required to execute requests from clients. A peer-to-peer (P2P) system is another type of information system which is composed of a huge number of peer computers where various types of applications are autonomously performed. In this paper, we consider a P2P system with data transfer application like the file transfer protocol (FTP). A computer consumes the electric power to transfer a file to another computer depending on the bandwidth. We discuss a model for power consumption of data transfer applications. A client peer has to find a server peer in a set of server peers which holds a file so that the power consumption of the server is reduced. We discuss algorithms to find a server peer which transfers file in a P2P overlay network.

Systems using peer-to-peer (P2P) overlay networks are getting a central position in information systems. P2P systems are in nature fully distributed, with no centralized coordinator and each peer is autonomous. Each peer has to obtain information on resource objects like databases distributed in networks by itself through communicating with its acquaintance peers. Since the P2P system is scalable, it takes time to propagate information on objects to peers in the network. A peer might be faulty and a peer holding an object might lose the object. Thus, an acquaintance peer might hold obsolete, even incorrect information on objects. It is critical to discuss how a peer can trust each acquaintance peer. We discuss two types of trustworthiness, subjective and objective ones. The subjective trustworthiness on an acquaintance peer is obtained by directly interacting with the acquaintance peer. On the other hand, the objective trustworthiness on an acquaintance peer is obtained by collecting the subjective trustworthiness of other peers. The subjective and objective types of trustworthiness might be different. If a peer is confident of its trustworthiness opinion, the peer takes the subjective trustworthiness. Otherwise, the peer rather takes the objective one. We discuss what type of trustworthiness a peer uses on an acquaintance peer based on the confidence.

In distributed peer-to-peer (P2P) applications, a group of multiple peer processes (peers) are required to cooperate with each other. Messages sent by peers have to be causally delivered. In this paper, we discuss a scalable group communication protocol for a group of multiple peers in P2P overlay networks. Due to the message overhead O(n) for the number n of peers, the vector clock cannot be used to causally deliver messages is a scalable group. On the other hand, the linear clock implies the message length O(1), but some pair of messages are unnecessarily ordered. Recently, more accurate physical clocks can be used with the GPS time server and network time protocol (NTP). In this paper, we consider a group where every member peer can use a physical clock which is synchronized with the time server in the network time protocol (NTP). Even if each physical clock is synchronized with a time server, every physical clock does not show the same accurate time. The accuracy of the physical clock depends on distance to the time server, traffic in a network and operating system. In this paper, we reduce the number of messages unnecessarily ordered by taking advantage of the linear time and physical time.

In peer-to-peer (P2P) applications, a group of multiple peer processes (peers) are required to cooperate with each other. In this paper, we discuss a heterogeneous hybridtime group communication (HHT) protocol which takes advantage of the linear time (LT) and physical time (PT) to causally order messages in a scalable heterogeneous group. It depends on accuracy of each physical clock and minimum delay time between a pair of peers how messages can be ordered. In this paper, we consider a heterogeneous type of group where the clock accuracy of each peer and the minimum delay time between every pair of peers are not the same. In group protocols, even if a pair of messages are ordered in the protocol, the messages may not be causally ordered. Thus, some messages are unnecessarily ordered in the protocols. In this paper, we show the number of messages to be unnecessarily ordered can be reduced in the HHT protocol. In a scalable group, it is not easy, maybe impossible for each peer to hold information on the clock accuracy and minimum delay time of every peer. In this paper, we newly consider a two-layered model of a heterogeneous group to reduce the information which each peer has to hold.

In peer-to-peer (P2P) applications, peers make an agreement on one opinion. Agreement procedures have to be so flexible that persons can change their opinions. We discuss a flexible agreement protocol of multiple peers by taking advantage of human behaviors in a fully unstructured P2P network. There are forward, backward, mining, and observation strategies for each peer to find a value to make an agreement. In the forward strategy, a peer shows a new value to the other peers. In the backward one, a peer backs to a previous round. In order to back to the previous round, values taken after the round have to be withdrawn. In the mining one, a peer tries to find a tuple of previous values named a cut which satisfies the agreement condition. In the observation one, a peer does not take a new value. Since each peer autonomously takes one of the strategies at each round, the peers might take inconsistent strategies. For example, only if every peer takes the mining strategy, the mining strategy can be adopted. The mining strategy is inconsistent with the others. The peers have to do negotiation with each other to take consistent strategies. We discuss how peers cooperate to take consistent strategies at each round.

A transaction is assigned with a purpose which is a collection of roles. Suppose a transaction T1writes an object o2after reading an object o1and then another transaction T2reads the object o2and writes an object o3. Here, data in the object o1might flow into o3via o2. Unless T2is granted a read access right of the object o1, illegal information flow occur. In order to prevent the illegal information flow, T1marks the object o2with the purpose of T1. T2cannot read o2unless the purpose of T2includes a read right of o1. In result, the throughput is degraded. Objects whose information may flow into an object o are source objects of o. An object is timed out if it takes some time units after the object is lastly written. While there occur no illegal information flow in our purpose marking (PM)protocol, transactions which imply illegal information flow are aborted. We evaluate the PM protocol in terms of how many transactions are aborted.

In peer-to-peer (P2P) overlay networks, source peers holding multimedia objects can transmit the multimedia objects to receiver peers. In this paper, we discuss parallel types of multi-source streaming (MSS) models where a receiver peer can receive primitive objects of a multimedia object from multiple source peers which can support enough. A multimedia object is realized in a sequence of primitive objects which are units of transmission at the P2P overlay layer. The receiver peer is required to receive primitive objects of the multimedia object with enough QoS, e.g. no primitive object loss. Multiple source peers in parallel send primitive objects to the receiver peer. In this paper, each primitive object is delivered to the receiver peer so that time to buffer the primitive object is reduced. In addition, primitive objects are duplicated by sending parity objects of the primitive objects. Here, the receiver peer can deliver every primitive object even if some primitive object is lost. We evaluate the model in terms of buffering time and delivery continuity in presence of network fault.

In peer-to-peer (P2P) applications, peers exchange their opinions with each other and make an agreement on one opinion. Agreement procedures have to be so flexible that persons can change their opinions, withdraw previous opinions under some constraints on the opinions, and use various types of agreement conditions like majority-condition in our society. We discuss a flexible agreement protocol of multiple peers by taking into account human behaviors in a fully unstructured peer-to-peer (P2P) system model. Each peer first notifies the other peers of its value. The peer in turn receives values from other peers. Unless the values satisfy the agreement condition, the peer can take another value. This is the forward strategy to make an agreement. In addition, we discuss other strategies, backward, mining, and observation ones in this paper. In the backward strategy, a peer backs to a previous round where the peer can take another value from previous values. In the mining one, a peer tries to find a tuple of previous values named a cut which satisfies the agreement condition. In the observation one, a peer does not take a new value and keeps the previous value. Each peer autonomously takes one of the strategies at each round. We discuss how peers cooperate to take one strategy at each round.

A process issues a transaction to manipulate objects. The transaction is assigned with a purpose which is a subfamily of roles granted to the process. Suppose a transaction T1writes an object o2after reading an object o1and then another transaction T2reads the object o2and writes an object o3. Here, data in the object o1might flow into the object o3via the object o2. Unless T2is granted a read access right of the object o1, illegal information flow occur. In order to prevent the illegal information flow, T1marks the object o2with the purpose of T1. T2cannot read o2unless the purpose of T2includes a read right of o1. In result, the throughput is degraded. Objects whose information may flow into an object o are source objects of o. If the source objects are written, a purpose mark on the object o is released. In addition, an object o might have some lifetime lambda when o's data has to be secure since the data is created. If it takes lambda time units since the object o is marked, the purpose mark is released. While there occur no illegal information flow in our purpose marking (PM) protocol, transactions which imply illegal information flow are aborted. We evaluate the PM protocol in terms of how many transactions are aborted.

In peer-to-peer (P2P) overlay networks, multimedia objects are distributed in peers and source peers transmit a multimedia object to receiver peers. In this paper, we would like to discuss how multiples source peers deliver multimedia objects to receiver peers in a streaming model so that enough quality of service (QoS) required is supported in presence of QoS change. A multimedia object is realized in a sequence of primitive objects which are units of transmission at the P2P overlay layer. The receiver peer is required to receive primitive objects of the multimedia object with enough QoS, e.g. no primitive object loss. We discuss types of multi-source streaming (MSS) models where a receiver peer can receive primitive objects of a multimedia object from multiple source peers which can support enough QoS.There are serial and parallel types of MSS models. In the serial MSS model, a receiver peer receives each primitive object from one of the source peers at a time. If a current source peer is expected to support lower QoS than required or not to support a succeeding primitive object, another source peer starts sending primitive objects. In the parallel MSS model, multiple source peers in parallel send primitive object units to the receiver peer. A primitive object is redundantly sent by multiple source peers or non-redundantly sent by a source peer. Here the throughput and reliability can be increased.

Information systems are composed of various types of computers interconnected in networks. In addition, information systems are being shifted from the traditional client-server model to the peer-to-peer (P2P) model. The P2P systems are scalable and fully distributed without any centralized coordination. It is getting more significant to discuss how to reduce the total electric power consumption of computers in information systems in addition to developing distributed algorithms to minimize the computation time. In this paper, we do not discuss the micro level like the hardware specification of each computer. We discuss a model to show the relation of the computation and the total power consumption of multiple peer computers to perform Web types of processes at macro level. We also discuss algorithms for allocating a process to a computer so that the deadline constraint is satisfied and the total power consumption is reduced.

In distributed applications, a group of multiple peer processes are required to cooperate with each other. Messages sent by peer processes have to be delivered in some order like causal order and total order. In this paper, we discuss a scalable group communication protocol for a group of multiple peers in peer-to-peer (P2P) overlay networks. Due to the message overhead O(n) for the number n of peer processes, the vector clock cannot be used to causally deliver messages. In this paper, we consider a group where every member peer process can use a physical clock which is synchronized with GPS time server in the network time protocol (NTP). The physical clock can be used to temporally order messages while the linear clock can be used to causally order messages. Even if each physical clock is synchronized with a time server, every physical clock does not show the same accurate time. The accuracy of the physical clock depends on distance, i.e. number of routers and traffic in a network. We discuss how the physical and linear types of clocks can be used to temporally and causally order messages in terms of delay time between peer processes and accuracy of each physical clock. In this paper, we discuss a group communication protocol to deliver messages by taking usage of physical clock and linear clock.

In this paper, we discuss a fully distributed unstructured P2P system where there is no centralized coordinator. Each process makes by itself a decision on an agreement value to realize some objectives by communicating with other processes in a group. Each process first takes a value v in a domain Di which is a collection of possible values and notifies the other processes of the value v. A process in turn receives values from other processes. Unless a set of the values satisfy an agreement condition, a process can take an other value v1, differently from the traditional consensus protocols. In human societies, people change their opinions in an agreement procedure. Types of precedent relations on values are discussed to show which value to be taken after a value v1is taken. An existentially precedent relation (v1rarr V2) shows a process pi can take a value v2after taking v1. In addition, values are ordered in the preferentially precedent relation (rarrP). A history of a process pi is a sequence of values which pi has taken. Some value may not be taken multiple times. Each value is characterized in terms of the maximum number of occurrences in a history. If a process pi finds some mistake on selecting a value at a previous round, the process would like to roll back to the previous round. A process takes a checkpoint at a round. The process pi can roll back to the checkpoint most recently taken. In this paper, we discuss the checkpoint mechanism in the coordination protocol. Based on the precedent relations and checkpoints, each process takes the most preferable value v1in the values which are changeable from current value v.

The role-based access control (RBAC) model is widely used to make information systems secure. Even if every access request is authorized in the roles, illegal information flow might occur as the well known confinement problem. In this paper, we discuss how to prevent illegal information flow to occur by synchronizing conflicting transactions in the RBAC model. We first define types of information flow relations, legal (LIF), illegal (IIF), and possibly illegal (PIF) ones R1= R2, R1rarr R2, and R1 rarr R2among a pair of role families R\ and Ri, respectively. Here, let T1and T2be a pair of transactions with role families R1and R2respectively. Suppose T1precedes T2in a schedule, i.e. for every pair of conflicting methods op\ and opi from T\ andT% respectively, op\ is performed prior to op2- Here, if the LIF relation R1= R2holds, no illegal information flow occur. If R1rarr R2, illegal information flow necessarily occur. R1rarr R2implies that illegal information flow might occur depending on in which order the transactions perform what methods.

There are many discussions on agreement protocols of multiple processes. However, they just aim at agreeing one value out of values shown by the processes. In human societies, agreement procedures are flexible, person can change their opinions and can use various type of agreement conditions. In this paper, we discuss a more flexible agreement protocol of multiple peer processes by taking account the human behaviour in social agreement procedures. Each process first takes a value v in a domain Di which is a collection of possible values and notifies the other processes of the value v. A process in turn receives values from other processes. Unless a set of the values from all the processes satisfy an agreement condition, a process can take another value v'. In order to model the social behaviour of people, we already discuss E- and P-precedent relations on values. In order to efficiently make an agreement, we have to discuss some coordination mechanism of multiple processes. A process pi would like to start over from a previous round if pi finds it to have selected an improper value. The process pi takes a checkpoint at a round. Then, the process pi can roll back to the previous checkpointed round. In addition, there are some constraints on a value. If a process says some value, the process cannot withdraw the value, i.e. cannot roll back. Based on the precedent relations and checkpoints, a group of processes efficiently make an agreement.

A peer-to-peer (P2P) overlay network is composed of peer processes. Various kinds of group-based applications like social network service (SNS) are developed and widely used in P2P overlay networks. Here, societies of peers are dynamically and autonomously organized to achieve some objectives in the centralized coordination. In this paper, we discuss a how multiple peers organize societies in the distributed coordination. Each peer is assumed to manage its local own objects in the role-based access control (RBAC) model in this paper. A role is a collection of access rights, i.e. permissions. A peer defines roles on its own objects and grants and revokes a role to and from another peer. Only a peer granted a role is allowed to make an access to the objects in access rights of the role. In this paper, we discuss how to securely share objects with multiple peers in a society. A society is defined to share local objects. A society is a collection of society objects and society roles to manipulate the society objects without being conscious of which local peer holds the objects and whether or not the peer is granted access rights. In a society, each member peer plays one or more than one society role. On joining a society, a peer first negotiates with the society an agreement on which roles the peer can support to the society and which society role the peer is granted. In this paper, we discuss how peers do the negotiation with each other to grant and revoke roles in a society.

In a wireless sensor-actuator network (WSAN) sensor nodes gather information on the physical world and can deliver messages with sensed values to only nearby nodes due to weak radio. Thus, messages sent by nodes might be lost due to not only collision but also noise. Messages are forwarded by sensor nodes to an actuator node. In the redundant data transmission (RT) protocol, a sensor node sends a message with not only its sensed value but also sensed values received from other sensor nodes. Even if a message with a sensed value v is lost, an actuator node can take the value v from a message sent by another sensor node. We evaluate the RT protocol in terms of how many number of sensed values an actuator node can receive in presence of message loss. We show that about 72 % of sensed values can be delivered to an actuator node even if 95 % of messages are lost due to noise and collision.

A sensor, actuator, and device network (SADN) is composed of sensor, actuator, and device nodes. Sensor nodes sense physical events and send sensed values of the events to actuator nodes. Actuator nodes decide on proper actions on receipt of sensed values and then issue requests of the actions to the device nodes. We propose an SADN model where sensor, actuator, and device functions are replicated in multiple nodes and each of sensor-actuator and actuator-device communication is realized in a communication protocol. A device node supports types of actions. For example, some action cannot be compensated by another action. An action has to be performed on only one device node. We classify actions into some types. For each type of action, we have to discuss a protocol for reliably performing the action on an device node in the presence of nodes and network faults. In this paper, we discuss how to perform actions on device nodes depending on types of the actions so that each action is surely performed on the required number of device nodes.

In traditional agreement protocols, multiple processes just aim at agreeing on one value out of values shown by the processes. In meetings of human societies, agreement procedures are often used and are so flexible that persons can change their opinions and can use various types of agreement conditions like majority-condition. We discuss a flexible agreement protocol of multiple peers by taking into account human behaviors in social agreement procedures in a fully unstructured peer-to-peer (P2P) model. We define a recoverable cut which is a satisfiable set of previous values which every peer can back. We discuss constraints on values which each peer can take at each round and a backward algorithm to find a satisfiable cut. If each peer takes a previous value in a recoverable cut, the peer can make an agreement. We discuss a protocol for multiple peers to find a recoverable cut.

In a peer-to-peer (P2P) system, multiple peer processes issue transactions to manipulate objects. A transaction issued by a process is assigned with a subfamily of roles named purpose, which are granted to the process. If a transaction reads an object o1and then writes another object o2, information in o1might flow into o2. Suppose a transaction T1with a purpose R1precedes another transaction T2with a purpose R2in a schedule. Here, if the legal information flow relation R1¿ R2holds, no illegal information flow occur. Otherwise, illegal information flow might occur. In order to prevent illegal information flow, if T1writes an object o, the object o is marked with the purpose R1. The transaction T2is allowed to read the object o if R1¿ R2. Even if T1commits, the object o isstill marked with the purpose R1. Hence, every transaction T3with such a purpose R3that R1does not legally flow into R3is aborted if T3reads the object o. In result, the throughput is degraded. In this paper, we discuss how to release purpose marks on objects. Objects whose information may flow into an object o are source objects of o. If the source objects are changed, a purpose mark on the object o is released. In addition, an object o might have some lifetime when o's data has to be secure since the data iscreated. If it takes time since the object o is marked, the purpose mark is released. While there occur no illegal information flow in our purpose marking (PM) protocol, transactions which imply illegal information flow are aborted. We evaluate the PM protocol in terms of how many transactions are aborted.

In information systems, multiple transactions issued by subjects manipulate objects in a conflicting way. Conflicting access requests from multiple transactions have to be serialized. There are various ways to order multiple access requests like FIFO and timestamp ordering (TO) schemes. In scheduling ways in operating systems and others, highly prioritized processes are performed before lower ones. In this paper, we postulate that an access request from a more significant subject should be performed prior to another access request from a less significant subject. We define the significance of subjects in terms of roles assigned to subjects and authorization of roles. In the role-based access control model, a role is a set of access rights. Only a subject granted a role can issue an access request in the role. Here, there are two views, subject-oriented (SO) and purpose-oriented (PO) to order transactions. A method issued by a more significant subject should be performed before another conflicting method issued by a less significant subject in the SO view. A transaction issued by a subject is associated with a purpose, i.e. a subset of roles granted to the subject. A method with a more significant purpose should be performed before another method with a less significant purpose in the PO view. In this paper, we discuss how conflicting methods are ordered in a unique SO and PO (SPO) view

In a wireless sensor and actuator network (WSAN), a group of sensor nodes, actuators, and actuation devices are geographically distributed and linked by wireless networks. Sensor nodes gather information for an event occurring in the physical world and send them to actuators. Actuators perform appropriate actions on actuation devices by making a decision on receipt of sensed values from sensor nodes. Sensor nodes are low cost, low powered devices with limited energy, computation, and wireless communication capabilities. Messages may be lost due to collision and noise and sensor nodes may be faulty. Here, nodes are required to reliably communicate with each other in realtime manner. In order to achieve the reliability, we discuss the multi-actuator/multi-sensor (MAMS) model where each sensor node sends sensed values to multiple actuators and each actuator receives sensed values from multiple sensor nodes in an event area. Even if messages are lost and sensor nodes are faulty, actuators can surely receive sensed values. An actuator makes a decision on what actions to be performed in what order. We discuss a semi-passive coordination (SPC) protocol for multiple actuators to communicate with multiple sensor nodes and perform actions. We evaluate the SPC protocol in terms of the number of messages exchanged among actuators

In peer-to-peer (P2P) applications like computer supported cooperative work (CSCW), multiple peer processes are required to cooperate to make a global decision, e.g. fix a meeting schedule of multiple persons. We discuss how multiple peer processes make a decision to achieve some objectives in a peer-to-peer (P2P) overlay network. Here, every process is assumed to be peer and autonomous. That is, there is no centralized coordination. A domain of a process is a collection of possible values which the process can take. Each process first takes a value v in its domain and notifies the other processes of the value v. A process can change the value with another value on receipt of values from other processes. However, a process can take only some value depending on the value v. For example, a process may abort after notifying commit but cannot commit after notifying abort in the commitment control. An existentially (E)-precedent relation shows what values a process can take after taking a value. In addition, a process takes a more preferable value if the process can take one of multiple values. Thus, values are ordered in the preferentially (P)-precedent relation. Based on the E- and P- precedent relations, each process takes the most preferable one in the values which can be changed from the current value v. In this paper, we discuss how every process makes an agreement on a value while each process can change the value according to the relations. In this paper, we discuss a coordination protocol in a type of heterogeneous system where every pair of processes have different E-precedent relation and P-precedent relation on the same domain. Each process learns a part of the precedent relations of another process through exchanging values.

In information systems, processes have to be scheduled to share a limited amount of resource objects like memory and CPU with other processes. In database systems, conflicting access requests from multiple transactions have to be serialized. There are various ways to order multiple conflicting access requests like FIFO and timestamp ordering (TO) schemes. For example, an object is exclusively used by each transaction in locking protocols. In scheduling algorithms, highly prioritized processes are performed before lower ones. Thus, it is critical to discuss which access request should be performed before another. In this paper, we postulate that an access request issued by a more significant subject for a more significant purpose should be performed prior to less significant access requests. In the role-based access control model, a role is a set of access rights, which shows a job function in an enterprise. A subject is first granted a role from another subject. Only a subject granted a role can issue an access request in the role. In this paper, the significancy of subject and purpose is defined in terms of roles and authorization relations. Here, there are two views, subject-oriented (SO) and purpose-oriented (PO) views to order transactions. A method issued by a subject with more significant roles should be performed before another conflicting method issued by a less significant subject in the SO view. A transaction issued by a subject is associated with a subset of roles granted to the subject, which is named purpose. A method with a more significant purpose should be performed before another method in the PO view. In this paper, we discuss how to combine the SO and PO views into a unique SPO view to order conflicting access requests.

Servers can be fault-tolerant through replication and checkpointing technologies in the client server model. However, application programs cannot be performed and servers might block in the two-phase commitment protocol due to the client fault. In this paper, we discuss the transactional agent model to make application programs fault-tolerant by taking advantage of mobile agent technologies where a program can move from a computer to another computer in networks. Here, an application program on a faulty computer can be performed on another operational computer by moving the program. A transactional agent moves to computers where objects are locally manipulated. Objects manipulated have to be held until a transactional agent terminates. Some sibling computers which the transactional gent has visited might be faulty before the transactional agent terminates. The transactional agent has to detect faulty sibling computers and makes a decision on whether it commits/aborts or continues the computation by skipping the faulty computers depending on the commitment condition. For example, a transactional agent has to abort in the atomic commitment if a sibling computer is faulty. A transactional agent can just drop a faulty sibling computer in the at-least-one commitment. We evaluate the transactional agent model in terms of how long it takes for the transactional agent to treat faulty sibling computers .

In distributed applications like computer supported cooperative work (CSCW), multiple peer processes are required to cooperate to make a global decision, e.g. fix a date for a meeting of multiple persons. We discuss how multiple peer processes make a decision to achieve some objectives in a peer-to-peer (P2P) overlay network. Here, every process is assumed to be peer and autonomous. A domain of a process is a collection of possible values which the process can take. Each process first takes a value v in its domain and notifies the other processes of the value v. A process can change the value to another value on receipt of values from other processes. However, a process may not take every value in the domain and can take only some value depending on the value v. For example, a process may abort after notifying commit but cannot commit after abort in the commitment control. An existentially dominant relation shows what values a process can take after taking a value. In addition, values are also ordered in the preferential relation. Based on the existential and preferential relations, each process takes the most preferable value in the domain, which is dominantly preceded by the value v. In this paper, we discuss how every process makes an agreement on a tuple of values while each process can change the value according to the existential and preferential relations. In this paper, we discuss a coordination protocol in a type of heterogeneous system where every pair of processes have the same domain but may have different existential and preferential relations. Each process learns a part of the relations of another process through exchanging values

We discuss a fully distributed peer-to-peer (P2P) overlay network where each peer is autonomous and accesses to objects distributed in peers by taking advantage of acquaintances. An acquaintance peer of a peer p is a peer about whose service the peer p knows and with which the peer p can directly communicate. The satisfiability is defined to how satisfiable a requesting peer is for each access request by newly taking into account the authorization. The trustworthiness of each acquaintance is calculated by aggregating the satisfiability obtained through each interaction with the acquaintance. If acquaintance information is changed, the change is propagated to peers through acquaintances. Due to the limited size of memory, the peer throws away information of less trustworthy acquaintances to make space to store new acquaintance information. The trustworthiness of acquaintance of each peer is propagated in a peer-by-peer way while some acquaintance information is recorded in a peer. We discuss how to calculate the trustworthiness, satisfiability, and exchange the acquaintance information then with acquaintances. We evaluate how the trustworthiness of acquaintance is changing through interactions among peers and the hit ratio can be improved for a detection request

In formation systems have to be kept consistent and secure in presence of multiple conflicting transactions and security threats. The role-based access control (RBAC) model is widely used to make systems secure. Here, a subject s is allowed to issue a method op to an object o only if an access right ho; opi is included in the roles granted to the subject s. Even if every access request issued by every subject is authorized in the roles, illegal information flow might occur as well known confinement problem. In this paper, we newly define a legal information flow (LIF) relation (R1⇒ R2) among a pair of role families R1and R2to prevent illegal information flow. Here, the relation R1⇒ R2shows that no illegal information flow occur if a transaction T1 with a role family R1 is performed prior to another transaction T2with R2. In addition, we discuss an illegal information flow (IIF) relation R1↦ R2, i.e. illegal information flow necessarily occur if every transaction T1with R1is performed before T2with R2. The more significant transaction, the more prior performed. We discuss a legal information flow (LIF) scheduler to synchronize transactions so as to prevent illegal information flow and to serialize conflicting methods from multiple transactions in terms of significancy and information flow relation of roles families.

In a wireless sensor-actuator network (WSAN), sensors send sensed values to actuators in a wireless channel. Messages sent by nodes might be lost due to collision and noise if multiple nodes simultaneously send the messages. If a message is detected to be lost, the source node retransmits the message to the destination nodes. It takes at least three rounds to detect and retransmit a lost message since the source node sends the message. The more number of messages are retransmitted, the more number of messages collide. We have to reduce the number of messages retransmitted to realize the efficient data transmission. We newly propose a redundant data transmission (RT) protocol where a sensed value is redundantly carried to the destination node by multiple messages. Even if a message is lost, the destination node can receive data in the lost message if the destination node receives other redundant messages. In addition, we have to reduce the energy consumption of a sensor node. A sensor node mainly consumes the energy to send and receive messages. Even if an event occurs, only some number of sensor nodes sensing the event send the sensed values to reduce the total energy consumption. We discuss an energy-efficient data transmission protocol. We evaluate the RT protocol compared with the CSMA protocol in terms of how much sensing data a node can receive in presence of messages loss

A large number of peer processes are distributed in a peer-to-peer (P2P) overlay network. It is difficult, maybe impossible for a peer to perceive the membership and location of every resource object due to the scalability and openness of a P2P network. In this article, we discuss a fully distributed P2P system where there is no centralized controller. Each peer has to obtain service information from its acquaintance peers and also send its service information to the acquaintance peers. An acquaintance peer of a peer p is a peer about whose service the peer p knows and with which the peer p can directly communicate in an overlay network. Some acquaintance peer might hold obsolete service information and might be faulty. Each peer has to find a more trustworthy one among acquaintance peers. There are many discussions on how to detect peers that hold a target object. However, a peer cannot manipulate an object without being granted access rights (permissions). In addition to detecting what peers hold a target object, we have to find peers granted access rights to manipulate the target object. The trustworthiness of each acquaintance is defined in terms of the satisfiability and ranking factor in this article. The satisfiability of an acquaintance peer shows how much each peer can trust the acquaintance peer through direct communication to not only detect target objects but also obtain their access rights. On the other hand, the ranking factor of an acquaintance peer indicates how much the acquaintance peer is trusted only by trustworthy acquaintance peers which is different from the traditional reputation concept. We evaluate how the trustworthiness of an acquaintance peer is changed through interactions among peers in a detection algorithm.

Service supported by an object is modeled to be a set of methods and quality of service (QoS). In addition to finding a peer which holds a target object, it is critical to discuss what peer is allowed to manipulate the target object in what method. In this paper, we take an acquaintance approach to finding and manipulating objects in P2P overlay networks. An acquaintance peer of a peer p is a peer whose service the peer p know and with which the peer p can directly communicate. If an access request is issued, a peer has to find where a target object exists, how to obtain an access right on the target object, and how to manipulate the target object. In this paper, we discuss ways to obtain results which satisfying an access requests by the cooperation of acquaintances. Acquaintance peers of a peer p may knows different information on target peers since it takes time to propagate change information of the target peers and peers may be faulty. Here, it is critical to discuss how much a peer can trust each acquaintance. We define the trustworthiness of an acquaintance peer in terms of the acquaintance relations among the peers.

A concept of role is widely and significantly used to design and implement secure information systems. A role shows a job function in an enterprise. In a role-based access control (RBAC) model, a role is a set of access rights. A subject doing a job is assigned with a role showing the job. In addition, a subject can play multiple roles in an enterprise. In addition to keeping systems secure, objects have to be consistent in presence of multiple conflicting transactions. Traditional locking protocols and timestamp ordering (TO) schedulers are based on principles "first-comer-winner" and "timestamp order" to make multiple conflicting transactions serializable, respectively. Since a transaction is associated with a collection of roles, we discuss which collection of roles is more significant than others. We discuss a role ordering (RO) scheduler so that multiple conflicting transactions are serializable in the significant dominant relation of roles.

In a peer-to-peer (P2P) overlay network, a large number and various types of peers are cooperating by exchanging multimedia contents. Here, multimedia streaming is a key technology to realize multimedia applications. A multimedia content is distributed to multiple peers and these peers named contents peers can support other leaf peers with the content. Multimedia contents are required to be efficiently delivered to processes in a real-time manner. Some contents peer may not send packets at a required rate and a communication channel may not support enough quality of service (QoS). In this paper, we newly discuss a heterogeneous asynchronous multi-source streaming (HAMS) model where multiple contents peers in parallel transmit packets of a multimedia content to a requesting leaf peer to realize highly reliability and scalability.

In a wireless sensor and actor network (WSAN), a group of sensors and actors are geographically distributed and linked by wireless networks. Sensors gather information sensed for an event in the physical world and send them to actors. Actors perform appropriate actions on actuation devices by making a decision on receipt of sensed information from sensors. Sensors are low cost, low powered devices with limited energy, computation, and wireless communication capabilities. Sensors may not only stop by fault but also suffer from arbitrary faults. Furthermore, wireless communication is less reliable due to noise and shortage of power of sensors. Reliable real time communication among sensors, actors, and actuation devices, is required in WSAN applications. We newly propose a multi-actor/multi-sensor (MAMS) model. In addition, multiple actors may perform actions on receipt of sensed information. Multiple redundant execution of an action on each device has to be prevented and conflicting actions on each device from multiple actors have to be serialized. In this paper, we discuss how to make WSAN reliable and available and how to reliably and non-redundantly perform actions with realtime constraints.

A transactional agent is a mobile agent to manipulate objects distributed on computers with some commitment condition like atomic commitment. Computers may stop by fault. In the client-server model, servers can be fault-tolerant according to replication and checkpointing technologies. However, an application program cannot be performed if a client computer is faulty. A transactional agent can move to another operational computer if some destination computer to which the agent to move is faulty. In this paper, we discuss how a program reliably manipulating objects can be realized in a mobile agent in presence of computer faults.

An object is a unit of resource distributed in a peer-to-peer (P2P) overlay network. Service supported by an object is modeled to be a set of methods. It is critical to discuss which peer can manipulate an object in which method, i.e. only a peer granted an access right (permission) can manipulate an object. First, an application has to find target peers which can manipulate a target object. In this paper, we take an acquaintance approach to finding and manipulating objects. Here, it is critical to discuss how much a peer trusts each acquaintance peer. We define the trustworthiness of each acquaintance peer in terms of how satisfiable a response from the peer is for each access request.

Multimedia contents are distributed to peers in various ways in peer-to-peer (P2P) overlay networks. A peer which holds a content, even a part of a content can provide other peers with the content. Multimedia streaming is more significant in multimedia applications than downloading ways in Internet applications. We discuss how to support peers with multimedia streaming service by using multiple contents peers. In our distributed multi-source streaming model, a collection of multiple contents peers in parallel transmit packets of a multimedia content to a requesting leaf peer to realize the reliability and scalability without any centralized controller. Even if some peer stops by fault and is degraded in performance and packets are lost and delayed in networks, a requesting leaf peer receives every data of a content at the required rate. We discuss a pair of flooding-based protocols, distributed and tree-based coordination protocols DCoP and TCoP, to synchronize multiple contents peers to reliably and efficiently deliver packets to a requesting peer. A peer can be redundantly selected by multiple peers in DCoP but it taken by at most one peer in TCoP. We evaluate the protocols in terms of how long it takes and how many messages are transmitted to synchronize multiple contents peers

A multimedia content are distributed to peers and a contents peer which holds contents can provide other peers with the contents in peer-to-peer (P2P) overlay networks. Multimedia streaming is more significant than downloading ways in multimedia applications from security and economical reasons. We discuss distributed multi-source streaming models. Here, a collection of multiple contents peers in parallel transmit packets of a multimedia content to a leaf peer to realize the reliability and scalability. Even if not only some number of peers stop by fault and are degraded in performance but also some number of packets are lost and delayed, a leaf peer can receive content at the required rate. We discuss how to allocate packets to each contents peer. We discuss a gossip-based, DAG-based coordination protocol (DCoP) to synchronize multiple contents peers. We evaluate DCoP in terms of how long it takes and how many messages are transmitted to synchronize multiple contents peers

In a wireless sensor and actor network (WSAN), a group of sensors, actors, and actuation devices are geographically distributed and linked by wireless networks. Sensors gather information for an event occurring in the physical world and send them to actors. Actors can perform appropriate actions on actuation devices by making a decision on receipt of sensed information from sensors. Sensors are low cost, low powered devices with limited energy, computation, and wireless communication capabilities. Sensors may not only stop by fault but also suffer from arbitrary faults. Furthermore, wireless communication is less reliable due to noise and shortage of power of sensors. Reliable realtime communication among sensors, actors, and actuation devices is required in WSAN applications. In order to realize the reliability and realtimeness, we newly propose a multi-actor/multi-sensor (MAMS) model where each sensor sends sensed information to multiple actors and each actor receives sensed information from multiple sensors in an event area. Actors are required to causally/totally order events from multiple sensors and actions on actuation devices. In addition, multiple actors may perform actions on receipt of sensed information. Multiple redundant executions of an action on each device have to be prevented and conflicting actions on each device from multiple actors have to be serialized. In this paper, we discuss how to realize reliable, ordered delivery of sensed information to actors from sensors on the basis of global time and how to reliably and non-redundantly perform actions with realtime constraints

A transactional agent is a mobile agent to manipulate objects distributed on computers with some type of commitment condition. For example, a transactional agent commits only if at least one object could be successfully manipulated in the at-least-one commitment condition. Computers may stop by fault while networks are assumed to be reliable. In the client-server model, servers can be fault-tolerant according to traditional replication and checkpointing technologies. However, an application program cannot be performed if a client computer is faulty. An application program can be performed on another operational computer even if a computer is faulty in the transactional agent model. For example, a transactional agent can move to another operational computer if some destination computer where the agent to move is faulty. There are kinds of faulty computers for a transactional agent, current, destination, and sibling computers where a transactional agent now exist, will move, and has visited, respectively. We discuss how the transactional agent can be tolerant of the types of faults. We show how a program reliably manipulating objects can be realized in a mobile agent in presence of computer faults

Role-based access control (RAC) models are widely used, to keep systems secure. In order to keep systems consistent, locking protocols and timestamp ordering schedulers are adopted, which, are based on principles "first-comer-winner" and "timestamp order'' to make multiple conflicting transactions serializable, respectively. In this paper, we discuss concurrency control algorithms based, on the role concept. We first define a. significantly dominant relation on roles showing which one of a pair of roles is more significant than the other role. We discuss a role ordering (RO) scheduler based, on the role concept. We evaluate the RO scheduler compared with the two-phase locking (2PL) protocol in terms of throughput and waiting time.

A peer-to-peer (P2P) network is composed of large number of peer computers which are cooperating by exchanging messages in networks. In multimedia streaming applications, multimedia data is required to be efficiently and reliably delivered to processes in a real-time manner. In order to support scalable and reliable multimedia streaming service, we newly discuss a multi-source streaming way where multiple source peers transmit packets of multimedia content to each peer. Each of the source peers transmits different packets with some redundancy to the peer in order to be tolerant of some number of faulty peers and packets lost.

Large number of peer processes distributed in various types of networks are cooperating to achieve some objectives. The vector clock cannot be adopted to a scalable group due to communication and computation overheads. In order to reduce the overhead, we discuss a group where all the processes are synchronized by physical clocks. Here, a subgroup of processes are interconnected in a local or personal area network. Clocks of computers in every local subgroup are synchronized by using the network time protocol (NTP) with a GPS time server. The maximum time difference among clocks of computers depends on types of operating systems of the computers. Every message is given physical time when a process sends the message. We discuss a global clock group (GCG) protocol where messages are causally ordered by using the time physical stamps with maximum time difference and delay time. We evaluate the protocol in terms of the number of messages ordered compared with the vector clock.

In this paper, we discuss a hierarchical group protocol aiming at reducing communication and computation supporting for realizing a scalable group. In traditional hierarchical groups, a pair of parent and child subgroups communicate with one another through a gateway process. A gateway process implies performance bottleneck and single point of failure. In order to increase the throughput and reliability of inter-subgroup communication, messages are in parallel transmitted in a striping way through multiple channels between the subgroups. We discuss how to design a hierarchical group for realizing high-performance multimedia communication among large number of peer processes.

Objects are concurrently manipulated through method issued by multiple transactions in object-based systems. Objects are locked in modes of the methods. Lock modes are defined based on the conflicting relation of method. We define a partially ordered relation on lock modes showing which modes are stronger than others. We newly propose a moderate locking (ML) protocol where an object is initially locked in a weaker mode than a mode intrinsic to the method and the lock mode is finally escalated to the method mode. The weaker the initial mode is, the higher concurrency is obtained but the more frequently deadlocks occur. The ML protocol is evaluated in terms of throughput and number of deadlocked transactions compared with the pessimistic and optimistic protocols.

A transactional agent is a mobile agent which manipulates objects distributed in computers. A transactional agent is composed of routing, manipulation, and commitment agents. A way to move to computers is decided in the routing agent. Objects in each computer are manipulated in a manipulation agent. After visiting computers, a transactional agent makes a destination on commitment by using the commitment agent. In addition, objects obtained from a computer in the manipulation agent have to be delivered to other computers where the transactional agent is performed. We discuss a model of transactional agent and implementation of a transactional agent on database servers and evaluate the transactional agents.

Multimedia objects are distributed in peer-to-peer (P2P) overlay networks since objects are cached, downloaded, and personalized in peer computers (peers). An application has to find target peers which can support enough quality of service of objects. We discuss a new type of flooding algorithm to find target peers based on charge and acquaintance concepts so that areas in networks where target peers are expected to exist are more deeply searched. In addition, we discuss how peers can be granted access rights to manipulate the objects with help and cooperation of acquaintances. We evaluate the charge-based flooding algorithm compared with a TTL-based flooding algorithm in terms of the number of messages transmitted in networks.

In a peer-to-peer (P2P) overlay network, a large number and various types of peer processes are cooperating by using multimedia contents like movies. Multimedia streaming is a key technology to realize multimedia applications. Here, multimedia contents are required to be reliable and continuously delivered to processes in a realtime manner. In this paper, we newly discuss a heterogeneous asynchronous multi-source streaming (HAMS) model where multiple contents peers transmit packets of a multimedia content to a requesting leaf peer to increase the throughput, reliability, and scalability in P2P overlay networks.

Multimedia objects are distributed on peer computers (peers) in peer-to-peer (P2P) overlay networks. An application has to find target peers which can support enough quality of service (QoS) of multimedia objects. We discuss types of acquaintance relations of peers with respect to what objects each peer holds, can manipulate, and can grant access rights. We discuss a new type of flooding algorithm to find target peers based on charge and acquaintance concepts so that areas in networks where target peers are expected to exist are more deeply searched. We evaluate the charge-based flooding algorithm compared with a TTL-based flooding algorithm in terms of the number of messages transmitted in networks.

This paper discusses a locking protocol to prevent illegal information flow among objects in a role-based access control model. In this paper, we newly define a conflicting relation "a role R/sub 1/ conflicts with another role R/sub 2/" to show that illegal information flow may occur if a transaction with R/sub 1/ is performed before another transaction with R/sub 2/. Here, we newly introduce a role lock on an object to abort a transaction with R/sub 1/ if another transaction with R/sub 2/ had been already performed on the object. Role locks are not released even if transactions issuing the role locks commit. A role lock on an object can be released if information brought into the object got obsolete. We discuss how to release obsolete role locks.

We discuss a group protocol which supports applications with group communication service in change of QoS supported by networks and required by applications. An autonomic group protocol is realized by cooperation of multiple autonomous agents. Each agent autonomously takes a class of each protocol junction. Classes taken by an agent are required to be consistent with but might be different from the others. A group is composed of views. Each view is a subset of the agents where the agents autonomously take protocol classes consistent with each other. We make clear what combination of classes can be autonomously taken by agents in a view. We also present how to autonomously change retransmission ways.

Multimedia objects distributed in networks are concurrently manipulated by multiple transactions in applications like co-authoring systems. Not only state but also quality of service (QoS) of object is changed through methods. We define new types of conflicting and exclusive relations on methods to serialize transactions and to realize exclusive manipulation of an object by taking into account QoS change in addition to state change. We discuss a protocol for locking objects and serializing transactions with respect to the QoS-based relations.

A large-scale network like the Internet is composed of various types of communication channels. Here, each communication channel supports quality of service (QoS) which may be different from others. In group communication, each process sends a message to multiple processes while receiving messages from multiple processes. In addition, messages are required to be causally delivered. In order to realize the atomic and ordered delivery of messages, even if a process with enough QoS channel receives messages, the process has to wait to deliver the messages until another process with lower QoS channel receives the messages. Thus, multimedia data cannot be delivered to processes so as to satisfy the real-time constraint if a slower process is included in a group. In this paper, we discuss group communication protocols by which multimedia messages can be delivered to a process with same time constraint properties which satisfy QoS requirements of multimedia messages.

A transactional agent is a mobile agent which manipulates objects in one or more than one computer so as to satisfy some constraint like ACID. An agent creates a surrogate agent on a computer on leaving the computer. A surrogate holds objects manipulated by the agent until the agent terminates. The surrogate can recreate a new incarnation of the agent if the agent is faulty. Transactional agents for multiple database servers are implemented in Aglets. We evaluate the performance in terms of total access time compared with traditional client-server model.

Large number of peer processes distributed in various types of networks are cooperating to achieve some objectives. The vector clock cannot be adopted to a scalable group due to the message length O(n) for number n of processes. A group is composed of local subgroups in each of which processes are in a local or personal area network and which are interconnected in a wide-area network. Processes in local subgroups use physical and linear clocks while processes in a wide-area network adopt vector clock. We discuss how to causally deliver messages by using local synchronization mechanisms of each subgroup. We evaluate the protocol in terms of number of messages ordered.

The focus is on the mechanism that supports the provision of the adequate service level by relying on Web system architecture with multinode clustering. Replication of Web documents can improve both performance and reliability of the Web service. Server selection algorithms allow Web clients to select one of the replicated servers which is "close " to them and thereby minimize the response time of the Web service.

In group communications, multiple processes first establish a group and then each process sends a message to multiple processes while receiving messages from multiple processes in the group. In addition, messages are required to be causally/totally delivered to each process. Due to the limited computation and communication resource, processes cannot send and receive as messages as the processes would like. We newly propose a notification-based data transmission procedure with two-phase slow start (TPSS) to efficiently exchange multimedia messages in a group so as to satisfy QoS requirement. In TPSS, the transmission rate of a process is increased by transmitting redundant data so that no data is lost even if some packets are lost.

Distributed multimedia objects are concurrently manipulated by multiple transactions. We define new types of conflicting and exclusive relations on methods to serialize transactions and to realize exclusive manipulation of an object by taking into account change of QoS and state. We discuss a locking protocol for realizing mutual exclusion on objects and a scheduler for serializing transactions.

A large-scale network is composed of various types of communication channels. Here, each communication channel supports a different quality of service (QoS). In group communication, messages are required to be causally delivered. Even if a process supporting enough QoS receives messages, the process has to wait to deliver the messages until another process supporting lower QoS receives the messages in order to realize the atomic and ordered delivery of messages. Thus, multimedia data cannot be delivered to processes so as to satisfy the real-time constraint if a slower process is included in a group. We discuss group communication protocols by which multimedia messages can be delivered to a process with some time constraint properties which satisfy the QoS requirements of multimedia messages.

Multimedia objects distributed in networks are concurrently manipulated by multiple transactions in applications like co-authoring systems and distributed gaming systems. Not only state but also quality of service (QoS) of object are changed through methods. We define new types of conflicting and exclusive relations on methods to serialize transactions and to realize exclusive manipulation of an object by taking into account QoS change in addition to state change. We discuss a locking protocol for realizing mutual exclusion on objects and a scheduler for serializing transactions with respect to the QoS-based conflicting and exclusive relations

The group protocol has to support applications with enough quality (QoS) and types of service in change of QoS supported by the underlying network and QoS requirements of applications. A flexible group service is supported for applications by cooperation of multiple autonomous agents. A group protocol to coordinate the cooperation of the agents is composed of a collection of functions like retransmission and receipt confirmation. Each agent dynamically and autonomously takes a class of module for each protocol function in change of QoS supported by networks and required by application.

In distributed applications, a group of multiple processes are cooperating by exchanging multimedia objects. The objects am decomposed into a sequence of messages which are units of transmission in a network. Following the causality relation among multimedia objects, we define novel types of causally precedent relations among messages decomposed from objects transmitted in a group of multiple processes. Only messages required to be causally ordered are ordered in each process. This means the overhead to causally deliver multimedia objects in a group can be reduced.

In this paper we discuss a hierarchical group of processes to support hundreds of processes. Each subgroup has gateway processes which communicate with other subgroups. Processes and gateway processes in each subgroup are locally supported with causally ordered delivery of messages by some mechanism like vector clock. Even if messages are locally causally ordered in a subgroup, the messages may not be globally causally ordered in the group. We discuss how global causality of messages among a pair of processes is related with local causalities of the messages in subgroups. In addition, we present a protocol to causally deliver messages through subgroups that are hierarchically structured.

This paper discusses a communication model to exchange multimedia messages in a group of multiple processes. Quality of Service (QoS) required by an application has to be supported. In traditional communication protocols like TCP, a process can reliably deliver messages to one or more than one process. In the group communication, a process sends messages to multiple processes while receiving messages from multiple processes in a group. We propose a notification-based data transmission procedure to exchange multimedia messages with processes so as to satisfy the QoS requirement.

We discuss how to invoke a method on multiple object replicas in a quorum-based way. Suppose each instance of a method t on replicas of an object x invokes another method u on replicas in a quorum of an object y. Here, the method u is redundantly invoked multiple times on some replicas of the object y. If each instance of the method t issues a method u to its own quorum, more number of replicas are manipulated than the quorum number This is quorum expansion. We discuss a protocol to invoke methods on replicas in a nested manner without the redundant invocation and quorum expansion. We evaluate the protocol on how many replicas are manipulated and requests are issued.

Distributed applications are realized by cooperation of multiple objects. A state of an object depends on in what order the objects exchange request and response messages. In this paper, we newly define an object-based precedent relation of messages based on conflicting relation among requests. Here, only the messages to be ordered in the object-based system are ordered and the others are not ordered. We discuss a protocol which supports the object-based ordered delivery of request and response messages. Here, an object vector is newly proposed to order messages.

Distributed applications are realized by cooperation of objects. A state of the object depends on in what order the object exchanges request and response messages and computes requests. In this paper, we newly define a significantly precedent order of messages based on a conflicting relation among requests. The objects can be mutually consistent if the objects take messages in the significantly precedent order. We discuss a protocol which supports the significantly ordered delivery of request and response messages. Here, an object vector is newly proposed to significantly order messages.

In group protocols, larger computation and communication are consumed to causally order all messages transmitted in the network. Transactions in clients manipulate objects in servers by sending read and write requests to the servers. In this paper we define significant messages, which are to be ordered at the application level, by using a conflicting relation among the transactions. We newly propose an object vector to causally order only the significant messages. The scheme of the object vector is invariant in the change of the group membership. We also show a TBCO (transaction-based causally ordered) protocol adopting the object vector; by which the number of messages to be causally ordered are reduced.

In group communications, larger computation and communication overhead are considered to causally order all the messages transmitted in the network. Transactions in clients manipulate objects in servers by sending read and write requests to the servers. In this paper, we define significant messages by using the relation among the transactions. We newly propose an object vector to causally order only the significant messages. The scheme of the object vector is invariant in the change of the group membership. We also show a TBCO (transaction-based causally ordered) protocol which adopts the object vector, by which the number of messages to be causally ordered are reduced.