Server load balancers (LBs) are critical components of interactive services, routing client requests to servers in a pool. LBs improve service performance and increase availability by spreading the request load evenly across servers.

The current pollution measurement methodology is coarse-grained where the pollution measurements are spatiotemporally few and far in-between. Our vision is to provide broadly accessible, fine-grained pollution information to a variety of end-users, and in turn, allow them to make better informed decisions using a new, more accurate information stream. To this end, this study proposes a new neural network model to estimate Carbon Monoxide (CO) concentrations outside vehicle from crowd-sourced CO measurements inside vehicles measured using mobile devices (dosimeters). End-users can benefit from the fine-grained pollution information generated by this prediction model along with data from direct measurements. A neural network is used to model the dynamic relationship between the CO measurements inside and outside a moving vehicle. The resulting neural network model is then used to predict outside CO concentrations from CO measurements inside vehicles. Mobile CO dosimeters were used inside and outside vehicles to collect measurements used in training a neural network based regression model. For this regression task, a new neural network architecture was designed using Convolutional layers and Gated Recurrent Unit (GRU) layers. The results show that outside CO concentrations can be estimated from inside vehicle CO measurements with high accuracy. The proposed neural network model provides a promising new and novel source of fine-grained pollution information along with direct measurement streams.

“What would have happened, had I taken the alternative route?” This is a question that many drivers often ponder, hoping that their chosen route is the best. This paper proposes DoppelDriver, a system that attempts to answer such questions on non-chosen alternative routes to a given destination by determining actual times of arrival (ATAs) from participatory users on the non-chosen routes. DoppelDriver offers a direct, actual travel time comparison among route choices. With DoppelDriver's collection of actual travel time comparisons, users will be able to make strategic decisions on, and self-assessments of, their route choices. Also, we describe the potential usage and benefits of ex-post feedback (i.e. travel time on non-chosen routes) and how snapshots of travel time comparisons can be used to support strategic decision making on the choice of a route. Using real taxi GPS data, we investigate whether aggregating ATAs for road segments from other users mimics the ATA for the intended origin-to-destination route. Finally, we present our system design for a prototype implemented on the Android platform.

Distributed online social networks (DOSN) have been proposed as an alternative to centralized Online Social Networks (OSN). In contrast to centralized OSN, DOSNs do not have central repository of all user data, neither impose control regarding how users data will be accessed. Therefore, users can keep control of their private data and are not at the mercy of the social network providers. However, one of the main problems in DOSNs is how to efficiently disseminate social updates among peers. In our previous work, we proposed Social Caches for social updates dissemination in DOSN. However, the selection of social caches requires knowledge about the entire social graph. In this paper, we propose four fully distributed social cache selection algorithms, and evaluate their performance on five well known graphs. Using simulations we show that these algorithms perform almost as good as the centralized best known approximation algorithm would do. These distributed caching techniques can be used as a basis for various applications such as those that represent fusions of social and vehicular networks.

Location-based services are growing in popularity due to the ubiquity of smartphone users. The relevance of location-based query results is very important, especially for mobile phones with limited screen size. Location-based data frequently changes; this introduces challenges in indexing and ranking places. The growing popularity of mobile social networks, such as Twitter, FourSquare and Facebook Places, presents an opportunity to build better location-based services by leveraging user interactions on these networks. In this paper, we present SocialTelescope, a location-based service that automatically compiles, indexes and ranks locations, based on user interactions with locations in mobile social networks. We implemented our system as a location-based search engine that uses geo-tweets by Twitter users to learn about places. We evaluated the coverage and relevance of our system by comparing it against current state-of-the-art approaches including page-rank (Google Local Search), expert-based (Zagat) and user-review based (Yelp). Our results show that a crowd-sourced location-based service returns results that match those returned by current approaches in relevance, at a substantially lower cost.

Network congestion can be alleviated either by reducing demand (traffic control) or by increasing capacity (resource control). Unlike in traditional wired or other wireless counterparts, sensor network deployments provide elastic resource availability for satisfying the fidelity level required by applications. In many cases, using traffic control can violate fidelity requirements. Hence, we propose the use of resource control: increasing capacity by enabling more nodes to become active during periods of congestion. However, a naive approach to increase resources without a careful consideration of the type of congestion, traffic pattern, and network topology make the situation worse. In this paper, we present TARA, a topology-aware resource adaptation strategy to alleviate congestion. The core of TARA is our capacity analysis model, which can be used to estimate capacity of various topologies. Detailed performance results show that TARA can achieve data delivery rate and energy consumption that is close to an ideal offline resource control algorithm.

In this paper, we first attempt to formally define the resource control framework that adjusts the resource provisioning at the hotspot during congestion. In an effort to find the optimal resource control under the fidelity and energy constraints, we present a resource increase and decrease algorithm called Early Increase/Early Decrease (EIED) that tries to adjust the effective channel capacity quickly to the incoming traffic volume in an energy-efficient manner, thereby increasing the fidelity (or accuracy) level observed by the application during congestion. Under the framework of energy-constrained optimization, we prove this algorithm incurs the lowest overhead of energy consumption for the given fidelity level that is required by the application.

Many sensor applications are being developed that require the location of wireless devices, and localization schemes have been developed to meet this need. However, as location-based services become more prevalent, the localization infrastructure will become the target of malicious attacks. These attacks will not be conventional security threats, but rather threats that adversely affect the ability of localization schemes to provide trustworthy location information. This paper identifies a list of attacks that are unique to localization algorithms. Since these attacks are diverse in nature, and there may be many unforeseen attacks that can bypass traditional security countermeasures, it is desirable to alter the underlying localization algorithms to be robust to intentionally corrupted measurements. In this paper, we develop robust statistical methods to make localization attack-tolerant. We examine two broad classes of localization: triangulation and RF-based fingerprinting methods. For triangulation-based localization, we propose an adaptive least squares and least median squares position estimator that has the computational advantages of least squares in the absence of attacks and is capable of switching to a robust mode when being attacked. We introduce robustness to fingerprinting localization through the use of a median-based distance metric. Finally, we evaluate our robust localization schemes under different threat conditions.

In this paper, we analyze the node scheduling approach of topology control in the context of reliable packet delivery. In node scheduling, only a minimum set of nodes needed for routing purposes (usually determined by a minimum connected dominating set, MCDS) are kept active. However, a very low density resulting from switching off nodes can adversely affect the performance of data delivery due to three factors. First, our analysis shows that at low density, the average path length increases by a factor more than previously thought. Second, protocols such as the Hop-By-Hop Broadcast (HHB) reliability scheme (which relies on high network degree for optimum performance) suffer. Third, with limited buffers at nodes, the overhead is more pronounced to the extent of making the network unstable. Using probabilistic models, we derive the relationship between network density and overhead based on the above factors and find the density conditions for minimum power consumption. We also propose a, fully distributed and message-optimal node scheduling algorithm with a constant approximation bound based on the concept of Virtual Connected Dominating Sets. The scheme can asymptotically achieve optimal density conditions while adapting to different network parameters.

Congestion in ad hoc networks not only degrades throughput, but also wastes scarce energy due to a large number of retransmissions and packet drops. For efficient congestion control, an accurate and timely estimation of resource demands by measuring the network congestion level is necessary. Congestion level measurement in ad hoc networks is more difficult than in wired networks due to time-variant channel capacity, contention among neighboring nodes, and non-deterministic node scheduling. We propose a new congestion detection mechanism that quantifies the congestion level accurately and energy-efficiently at both a node-level (implemented at the MAC layer) and a flow-level (implemented at the routing layer) in ad hoc networks. For accurate congestion measurement, a set of metrics that decouple the measurement from various MAC protocol characteristics is defined. For energy-efficient congestion measurement, an asynchronous channel loading measurement scheme, called lazy measurement, which emulates synchronous measurement by using virtual channel sampling, is incorporated into the proposed scheme. Simulation results show that the proposed mechanism significantly cuts down the energy needed to measure congestion accurately, while maintaining the high level of accuracy needed for timely congestion control.

Angle of arrival (AOA) has previously been used for outdoor positioning in aircraft navigation and for services like E911. For indoor positioning, the best schemes to date rely either on extensive infrastructure, or on sampling of the signal strength on a dense grid, which is subject to changes in the environment, like furniture, elevators, or people. We present an indoor positioning architecture that does not require a signal strength map, simply requiring the placement of special VOR base stations (VORBA). While our incipient realization of the AOA using 802.11 uses a base station with a revolving directional antenna, a non mechanical implementation would yield comparable performance, even with quantized angles. Performance of positioning with VOR base stations is evaluated though experimentation, simulation, and theoretical analysis.

APS algorithms use the basic idea of distance vector routing to find positions in an ad hoc network using only a fraction of landmarks, for example GPS enabled nodes. All the nodes in the network are assumed to have the possibility of measuring: range, angle of arrival (AOA), orientation, or a combination of them. We give a lower bound for positioning error in a multihop network for a range/angle free algorithm, and examine the error characteristic of four classes of multihop APS algorithms under various condition, using theoretical analysis and simulations. Analysis of range/angle free, range based, angle based, and multimodal algorithms show a complex tradeoff between the capabilities used, the density of the network, ratio of landmarks, and the quality of the position obtained.

A fundamental issue in the design of a wireless sensor network is to devise mechanisms to make efficient use of its energy, and thus, extend its lifetime. Due to the importance of energy conservation, it is highly desirable to define the amount of energy each protocol can spend to perform its goal. Using this idea, we can associate a finite energy budget with each network activity, and ask this activity to achieve its best performance using only its budget. This should be considered a new paradigm to design algorithms for networks that are battery powered, specially for wireless sensor networks. In this paper, we present this new paradigm, and show how it can be used to construct the energy map of a wireless sensor network. Our goal is to construct the best energy map using only a defined amount of energy. Results show that we can approach these performance limits using the proposed finite energy budget model.

As cellular packet data services become widely deployed by the rollout of the networks such as general packet radio service (GPRS) and 3G cellular networks, packet-switched voice service such as voice over IP (VoIP) can soon be expected to be offered as an alternative to circuit-switched voice service. While circuit-switched voice offers better quality, packet-switched voice offers better resource utilization due to its multiplexing and compression techniques that can be used in packet radio principle. Therefore, the capability of having voice service over either circuit-switched network or packet-switched network provides a tradeoff between voice quality and bandwidth utilization. This paper proposes three schemes that allow users to toggle between circuit-switched voice and packet-switched voice based on the desired tradeoff. The basic idea is to allow ongoing voice traffic to alternate between circuit-switched network and packet- switched network using a 3-way calling mechanism in the SIP-enabled GSM/GPRS cellular network. We also show how these toggle schemes can be used depending on the cell load. Detailed performances of the proposed schemes are evaluated using simulation with realistic parameters of a cellular environment. The result shows our schemes can alleviate cell overloading as well as provide the smooth transition of an ongoing call between circuit-switched network and packet-switched network. Our schemes can be easily deployed in any cellular mobile network where circuit-switched network and packet-switched network coexist.

Max-min fair bandwidth allocation is desirable because it protects well-behaved flows while keeping network utilization high. Achieving and maintaining absolute global max-min fairness at finite flow granularity is improbable. Moreover, computing and communicating max-min fair rates has a significant overhead. In this paper, we define the notion of /spl epsiv/-fairness and derive the conditions for flow rates to be /spl epsiv/-fair. We describe a technique where max-min fairness is computed at the level of flow groups while /spl epsiv/-fairness is provided to individual flows. Using /spl epsiv/-fairness allows the network administrator to control the degree of fairness with an overhead proportional to the degree. An important implication of our technique is that traffic-pattern knowledge can be used to reduce overhead.

The core-stateless service architecture alleviates the scalability problems of the integrated service framework while maintaining its guaranteed service semantics. The admission control methods proposed to support core-stateless guaranteed services have significant drawbacks: We propose a scalable and robust distributed admission control architecture to support core-stateless guaranteed services. Our architecture maintains high network utilization while ensuring that resources are not over-allocated. In our architecture, admission control is performed at the ingress edge routers of a request on an edge-to-edge path basis. A token-passing mechanism is used as the resource management framework. The token helps in dynamic and fair division of bandwidth and allows completely distributed resource allocation on a link unless it is close to saturation. The edge routers co-operate to provide fault tolerance effectively acting as a resilient overlay network. Our admission control framework can support statistical guarantees and diffserv architecture's premium service as well. The resource management part of our architecture is well-suited to aid QoS routing algorithms. Analytical and simulation results are presented to show the effectiveness of our architecture.

Efficient bandwidth allocation strategy with simultaneous fulfillment of QoS requirement of a user in a mobile cellular network is still a critical and an important practical issue. We explore the problem of finding the reservation schedule that would minimize the amount of time for which bandwidth has to be allocated in a cell while meeting the QoS constraint. With the knowledge about the arrival and residence time distribution of a user in a cell, the above problem can be optimally solved using a dynamic programming based approach in polynomial time. To be able to use the solution, we provide a mechanism for constructing the arrival/residence time distribution based on the measurement of hand-off events in a cell. The above solution allows us to propose an optimal time based bandwidth reservation and call admission scheme. By being scalable and distributed, the proposed scheme justifies for practical implementation. Simulations results are also presented to show the effectiveness of the scheme to achieve the target QoS level and optimal bandwidth utilization.

Position information of individual nodes is useful in implementing functions such as routing and querying in ad-hoc networks. Deriving position information by using the capability of the nodes to measure time of arrival (TOA), time difference of arrival (TDOA), angle of arrival (AOA) and signal strength have been used to localize nodes relative to a frame of reference. The nodes in an ad-hoc network can have multiple capabilities and exploiting one or more of the capabilities can improve the quality of positioning. In this paper, we show how AOA capability of the nodes can be used to derive position information. We propose a method for all nodes to determine their orientation and position in an ad-hoc network where only a fraction of the nodes have positioning capabilities, under the assumption that each node has the AOA capability.

Sensor networks are meant for sensing and disseminating information about the environment they sense. The criticality of a sensed phenomenon determines its importance to the end user. Hence, data dissemination in a sensor network should be information aware. Such information awareness is essential firstly to disseminate critical information more reliably and secondly to consume network resources proportional to the criticality of information. In this paper, we describe a protocol called RelnForM to deliver packets at desired reliability at a proportionate communication cost. RelnForm sends multiple copies of each packet along multiple paths from source to sink, such that data is delivered at the desired reliability. It uses the concept of dynamic packet state in context of sensor networks, to control the number of paths required for the desired reliability, and does so using only local knowledge of channel error rates and topology. We show that for uniform unit disk graphs, the number of edge-disjoint paths between nodes is equal to the average node degree with very high probability. RelnForm utilizes this property in its randomized forwarding mechanism which results in use of all possible paths and efficient load balancing.

Trajectory based forwarding (TBF) is a novel methodto forward packets in a dense ad hoc network that makes it possible to route a packet along a predefined curve. It is a hybrid between source based routing and Cartesian forwarding in that the trajectory is set by the source, but the forwarding decision is based on the relationship to the trajectory rather than names of intermediate nodes. The fundamental aspects of TBF are: it decouples path naming from the actual path; it provides cheap path diversity; it trades off communication for computation. These aspects address the double scalability issue with respect to mobility rate and network size. In addition, TBF provides a common framework for many services such as: broadcasting, discovery, unicast, multicast and multipath routing in ad hoc networks. TBF requires that nodes know their position relative to a coordinate system. While a global coordinate system afforded by a system such as GPS would be ideal, approximate positioning methods provided by other algorithms are also usable.

We present our ongoing work on data quality problems in sensor networks. Specifically, we deal with the problems of outliers, missing information, and noise. We propose an approach for modeling and online learning of spatio-temporal correlations in sensor net-works. We utilize the learned correlations to discover outliers and recover missing information. We also propose a Bayesian approach for reducing the effect of noise on sensor data online.

It is an inherently challenging task to provide quality of service (QoS) to real time applications in cellular networks owing to their mobility-induced uncertainties. Here we consider an adaptive framework where services may undergo bandwidth degradation. In contrast to the existing schemes, we here provide a reservation states based approach to provision the QoS to the individual adaptive applications. By using information related to users' mobility, the proposed approach can assure probabilistic QoS to multiple level of degradation as maybe required by such adaptive services. Simulation results show the effectiveness of the scheme in realizing the target QoS.

Resource reservation - both just in time and advance in time - has been advocated as a means of guaranteeing quality of service (QoS) to mobile users. Often, advance bandwidth reservation results in low utilization due to bandwidth fragmentation wherein holes of unused bandwidth in time may be too small to allocate to any user yet the sum of several holes may be sufficient to satisfy the user. In this paper, we introduce and investigate the problem of bandwidth fragmentation and explore different methods of minimizing bandwidth fragmentation. One is a selective call rejection scheme and the other a multi-class bandwidth sharing scheme. Through an extensive set of simulation results, we show that our proposed schemes reduces bandwidth fragmentation.

To provide quality of service guarantees, resource reservation schemes have to maintain reservation states at the nodes along the path of a flow. Advance reservation schemes have to maintain these states for a long period of time. The loss of reservation state caused by node failures, makes advance reservation schemes highly susceptible to node failures. In this paper, we argue that a domain-by-domain reservation approach is inherently more fault tolerant than the traditional hop-by-hop approach. We propose a novel domain-based protocol for handling advance reservations. It requires support only at the edge routers and no changes are required at the core routers. We describe various factors which determine whether a flow's advance reservation is handled properly by a network in the event of router failures. We use simulations to show that our protocol is highly fault tolerant. Our protocol does not require any static partitioning of bandwidth between best effort flows and reserved flows and thus keeps network utilization high.

It is an inherently challenging task to provide quality of service (QoS) to real time multimedia users in a cellular network owing to their mobility induced uncertainty. We propose an implicit resource allocation strategy and justify the need of such a strategy toward an effective solution to QoS provisioning. The reservation scheme presented, based on this proposed strategy, aims at meeting the users' QoS requirement while maintaining a better utilization as compared to the conventional resource reservation schemes. Simulation results are presented which show the effectiveness of the scheme under a varying level of user mobility.

Many ad hoc network protocols and applications assume the knowledge of geographic location of nodes. The absolute location of each networked node is an assumed fact by most sensor networks which can then present the sensed information on a geographical map. Finding location without the aid of GPS in each node of an ad hoc network is important in cases where GPS is either not accessible, or not practical to use due to power, form factor or line of sight conditions. Location would also enable routing in sufficiently isotropic large networks, without the use of large routing tables. We are proposing APS - a distributed, hop by hop positioning algorithm, that works as an extension of both distance vector routing and GPS positioning in order to provide approximate location for all nodes in a network where only a limited fraction of nodes have self location capability.

IP multicast is an efficient point-to-multipoint distribution mechanism. However, there are a number of scenarios in which a reverse, multipoint-to-point aggregation mechanism is highly desirable, We introduce a programmable mechanism, called gathercast, to support the aggregation of packets without altering any of the routing or forwarding mechanisms of the Internet. Gathercast is based on an active services framework and can be deployed incrementally. It works well within the current IP multicast model. One of the aggregation mechanisms that we study is the combination of small packets using gathercast. Small packets constitute a large fraction of packets in today's Internet. Every packet requires a routing table lookup at every router and incurs the same performance cost irrespective of the size of the packet. Gathercast allows for car pooling, in which multiple small packets to the same destination are combined into a larger packet, thereby reducing the number of packets in the network. Web servers, in particular, are a leading cause of small packets. All the TCP ACKs generated by the clients have the same destination (the Web server), These ACKs are just 40 bytes in size, Simulation studies show 11% to 68% reduction in packet loss and 2% to 30% improvement in performance of busy Web servers when used in conjunction with gathercast.

As mobile computing devices become more and more popular mobile databases have started gaining popularity. An important feature of these database systems is their ability to allow optimistic replication of data by permitting disconnected mobile devices to perform local updates on replicated data. The fundamental problem in this approach is the reconciliation problem, i.e. the problem of serializing potentially conflicting updates performed by local transactions on disconnected clients on all copies of the database. We introduce a new algorithm that combines multiversion concurrency control schemes on a server with reconciliation of updates from disconnected clients. The scheme generalizes to multiversion systems, the single version optimistic method of reconciliation, in which client transactions are allowed to commit on the server iff data items in their read sets are not updated on the server after replication.

This paper considers the problem of handoff management in an Integrated Services Packet Network supporting mobile hosts. In such a network, links may become overloaded due to a high concentration of mobile hosts within a given cell. As a result, the required Quality of Service guarantees cannot be provided to allows in that cell. However, there exist adaptive applications which can operate over a wide range of available bandwidth. Thus, it may be possible to overcome the link overload condition by reducing the bandwidth of individual rows, which we call rate adaptation. The two most important properties of a rate adaptation scheme are its network overhead and its fairness property. Rate adaptation schemes which ensure certain fairness properties, such as maxmining optimality criteria, have very high network overhead. Therefore, these schemes are not suitable for a highly mobile environment where rate adaptation may have to be invoked frequently. In this paper, we investigate the tradeoff between the network overhead and the fairness property of rate adaptation in a mobile environment. We first characterize the fairness property of rate adaptation by several measurable parameters. We then describe two rate adaptation schemes, one of which has very low network overhead but is `unfair', while the other scheme is `fair' but has a very high network overhead. Finally, we propose a new rate adaptation scheme which reconciles the two conflicting properties. Results of simulation experiments comparing the performances of the three rate adaptation schemes are presented.

In mobile computing, factors such as add-on hardware components and heterogeneous networks result in an environment of changing resource constraints. An application in such a constrained environment must adapt to these changes so that available resources are properly utilized. We propose an architecture for exporting awareness of the mobile computing environment to an application. In this architecture, a change in the environment is modeled as an asynchronous event that includes information related to the change. Events are typed and are organized as an extensible class hierarchy so that they can be handled at different levels of abstraction according to the requirement of each application. We also compare two approaches to structure an adaptive application. One addresses the problem of incorporating adaptiveness into legacy applications, while the other considers the design of an application with adaptiveness in mind.

In mobile computing, factors such as add-on hardware components and heterogeneous networks result in an environment of changing resource constraints. An application in such a constrained environment must react to these changes so that available resources are properly utilized. In this paper, we propose a framework to build environment aware applications. The architecture is based on an event delivery mechanism that decouples event detection from delivery, giving the flexibility and extensibility that is necessary in a mobile computing environment. Information associated with the event is delivered as part of the event notification, while delivery latency is reduced by clever thread scheduling. We demonstrate the utility of this framework by structuring an environment aware networking subsystem around a prototype implementation. The performance of this implementation is competitive with current event delivery mechanisms such as the Unix signal.

This paper considers the support of real-time applications to mobile hosts in an integrated services packet network. We have proposed a service model for mobile hosts that can support adaptive applications which can withstand a wide range of available bandwidth, as well as applications which require mobility independent service guarantees. We describe an admission control scheme and a reservation protocol for implementing this service model. Our admission control scheme achieves high utilization of network resources.

With the advent of small portable computers and the technological advances in wireless communications, mobile wireless computing is likely to become very popular in the near future. Wireless links are slower and less reliable compared to wired links and are prone to loss of signal due to noise and fading. Furthermore, host mobility can give rise to periods of disconnection from the fixed network. The use of existing network protocols, which were developed mainly for the high bandwidth and faster wired links, with mobile computers thus gives rise to unique performance problems arising from host mobility and due to the characteristics of wireless medium. Indirect protocols can isolate mobility and wireless related problems using mobility support routers (MSRs) as intermediaries, which also provide backward compatibility with fixed network protocols. We present the implementation and performance evaluation of I-TCP, which is an indirect transport layer protocol for mobile wireless environments. Throughput comparison with regular (BSD) TCP shows that I-TCP performs significantly better in a wide range of conditions related to wireless losses and host mobility. We also describe the implementation and performance of I-TCP handoffs.

Organizing massive amount of data on wireless communication networks in order to provide fast and low power access to users equipped with palmtops, is a new challenge to the data management and telecommunication communities. Solutions must take under consideration the physical restrictions of low network bandwidth and limited battery life of palmtops. This paper proposes algorithms for multiplexing clustering and nonclustering indexes along with data on wireless networks. The power consumption and the latency for obtaining the required data are considered as the two basic performance criteria for all algorithms. First, this paper describes two algorithms namely, (1, m) indexing and Distributed Indexing, for multiplexing data and its clustering index. Second, an algorithm called Nonclustered Indexing is described for allocating static data and its corresponding nonclustered index. Then, the Nonclustered indexing algorithm is generalized to the case of multiple indexes. Finally, the proposed algorithms are analytically demonstrated to lead to significant improvement of battery life while retaining a low latency.

As mobile hosts move, they encounter changing network characteristics. Characteristics such as bandwidth, reliability can change drastically when a mobile host moves from indoor to outdoor environment and vice versa. A mobile host can find itself in a lossy environment due to external conditions such as interference and fading. Existing end-to-end protocols are designed with an assumption that packet loss is a random and rare event. However due to mobility of hosts, there can be transient yet significant packet loss. In this paper we consider the effect of mobility on unreliable protocols such as UDP. We provide a mechanism by which unnecessary loss of packets can be avoided by allowing transmission only under "good" link conditions.

This paper deals with multicasting in an internetwork with mobile hosts, particularly with regard to mobile-IP and distance vector multicast routing (DVMRP) protocols. When the source of a multicast datagram is a mobile host (MH), the datagram may not reach all group members to which the datagram is addressed, including other mobile hosts. When the source is a static host and the multicast group includes mobile hosts, a mobile group member may receive datagrams in one cell but not in another. Further when a MH enters a cell which contains no other member of the same group, the MH will experience a delay before it starts receiving datagrams addressed to that group. Mobility between campuses which result in a MH acquiring an additional unicast address, also has an effect on multicast routing. We propose enhancements to DVMRP executed at the mobility support routers (MSR) that ensure correct forwarding of multicast datagrams to and from mobile hosts. Our solutions do not require any change at hosts and routers unaware of mobility, i.e. the modifications are limited to MSRs and MHs. We also describe an implementation incorporating a subset of our proposals. Lastly, we show that alternate styles of multicasting or mobile networking, viz. link-state (MOSPF) and IETF proposal, will face similar problems and our proposed solutions are still valid in their context.

IP based solutions to accommodate mobile hosts within existing internetworks do not address the distinctive features of wireless mobile computing. IP-based transport protocols thus suffer from poor performance when a mobile host communicates with a host on the fixed network. This is caused by frequent disruptions in network layer connectivity due to i) mobility and ii) unreliable nature of the wireless link. We describe I-TCP, which is an indirect transport layer protocol for mobile hosts. I-TCP utilizes the resources of Mobility Support Routers (MSRs) to provide transport layer communication between mobile hosts and hosts on the fixed network. With I-TCP, the problems related to mobility and unreliability of wireless link are handled entirely within the wireless link; the TCP/IP software on the fixed hosts is not modified. Using I-TCP on our testbed, the throughput between a fixed host and a mobile host improved substantially in comparison to regular TCP.

Mobile intemetworking revolves around the premise that a mobile host (MH) will frequently attach to and detach from the Intemet, thus changing its addressfrequently.Location management in mobile intemetworking refers to the process of keeping track of the current addresses of MHs in the Intemet.Protocols proposedfor mobile intemetworking rely on a combination of searchesand updates for location management.Adrawback of these protocols is that they are not uniformly efficient over all possible call-to-mobility ratios (i.e., the relative frequency of searchesascompared to up- dates).Unimportant issue, therefore, ishowthe overall costs of location management can be reduced, regardless of the call-to-mobility ratio.The aim of this work is to explore a new and practical scheme for location management, with emphasis on network cost reduction.We makethe keyobservation thatwhile thepotential setof sources forthe MHmaybelmge, tiesetofsources thatagiven MH communicates most frequently with is very small.Based on this, we develop the concept of a working set o~hosts for the MH (similar to the working set concept in operating systems).An adaptive scheme of location management is proposed, that enables an MH to dynamically determine its working set, and trade-off routing and update costs in order to reduce the total cost.Comparative analysis of the proposed scheme, using simulation, shows its efficiency over a wide range of call-to-mobility ratios.

Distributed algorithms have hitherto been designed for networks with static hosts. A mobile host (MH) can connect to the network from different locations at different times. This paper presents an operational system model for explicitly incorporating the effects of host mobility and proposes a general principle for structuring efficient distributed algorithms in this model. This principle is used to redesign two classical algorithms for distributed mutual exclusion for the mobile environment. We then consider a problem introduced solely by host mobility viz, location management for groups of MHs, and propose the concept of group location as an efficient approach to tackle the problem. Lastly, we present a framework which enables host mobility to be decoupled from the design of a distributed algorithm per se, to varying degrees.>

We consider wireless broadcasting of data as a way of disseminating information to a massive number of users. Organizing and accessing information on wireless communication channels is different from the problem of organizing and accessing data on the disk. We describe two methods, (1,m) Indexing and Distributed Indexing, for organizing and accessing broadcast data. We demonstrate that the proposed algorithms lead to significant improvement of battery life, while retaining a low access time.

There is a strong trend toward integrating portable computers within existing data networks. Traditional network protocols were designed assuming a static view of network connectivity. A mobile host can connect to the network from different locations at different times. This has led to the emergence of a new set of problems with regard to addressing schemes and network protocols for accommodating mobility within existing networks. The authors present a protocol for delivering a multicast message exactly once to a group of mobile destinations. The protocol is based on a system model derived from an architecture developed by J. Ioannidis et al. (1991), for mobile internetworking, relying on mobile support stations within the fixed network to communicate with mobile hosts.>

The client-agent-server architecture, which serves as a flexible mechanism for many-to-many remote procedure calls, is discussed. The architecture provides a transparent view of the server to a client. The client accesses the server group through the agent while retaining only an abstract view of the server. The call primitives for the client, agent groups, hook construction and primitives (hooks are basic building block which a server group provides to the agent group), and server groups used in the architecture are described. Implementation issues are reviewed.>

The concurrency of transactions executing on atomic data types can be enhanced through the use of semantic information about operations defined on these types. Hitherto, commutativity of operations has been exploited to provide enchanced concurrency while avoiding cascading aborts. We have identified a property known as recoverability which can be used to decrease the delay involved in processing noncommuting operations while still avoiding cascading aborts. When an invoked operation is recoverable with respect to an uncommitted operation, the invoked operation can be executed by forcing a commit dependency between the invoked operation and the uncommitted operation; the transaction invoking the operation will not have to wait for the uncommitted operation to abort or commit. Further, this commit dependency only affects the order in which the operations should commit, if both commit; if either operation aborts, the other can still commit thus avoiding cascading aborts. To ensure the serializability of transactions, we force the recoverability relationship between transactions to be acyclic. Simulation studies, based on the model presented by Agrawal et al. [1], indicate that using recoverability, the turnaround time of transactions can be reduced. Further, our studies show enchancement in concurrency even when resource constraints are taken into consideration. The magnitude of enchancement is dependent on the resource contention; the lower the resource contention, the higher the improvement.

For next generation information systems, concurrency control mechanisms are required to handle high level abstract operations and to meet high throughput demands. The currently available single level concurrency control mechanisms for reads and writes are inadequate for future complex information systems. In this paper, we will present a new multilevel concurrency protocol that uses a semantics-based notion of conflict, which is weaker than commutativity, called recoverability. Further, operations are scheduled according to relative conflict, a conflict notion based on the structure of operations.Performance evaluation via extensive simulation studies show that with our multilevel concurrency control protocol, the performance improvement is significant when compared to that of a single level two-phase locking based concurrency control scheme or to that of a multilevel concurrency control scheme based on commutativity alone. Further, simulation studies show that our new multilevel concurrency control protocol performs better even with resource contention.

A method is discussed for synchronizing operations on objects when the operations are invoked by transactions. The technique, which is motivated by a desire to make use of possible concurrency in accessing objects, takes into consideration the granularity at which operations affect an object. A dynamic method is presented for determining the compatibility of an invoked operation with respect to operations in progress. In making decisions, it utilizes the state of the object, the semantics of the uncommitted operations, the actual parameters of the invoked operation, and the effect of the operations on the objects. One of the attractive features of this technique is that a single framework can be used to deal with the problem of synchronizing access to simple objects as well as compound objects, i.e. objects in which some components are themselves objects.>

The concurrency of transactions executing on atomic data types can be enhanced through the use of semantic information about operations defined on these types. Hitherto, commutativity of operations has been exploited to provide enhanced concurrency while avoiding cascading aborts. We have identified a property known as recoverability which can be used to decrease the delay involved in processing non-commuting operations while still avoiding cascading aborts. When an invoked operation is recoverable with respect to an uncommitted operation, the invoked operation can be executed by forcing a commit-dependency between the invoked operation and the uncommitted operation; the transaction invoking the operation will not have to wait for the uncommitted operation to abort or commit. Further, this commit dependency only affects the order in which the operations should commit, if both commit; if either operation aborts, the other can still commit thus avoiding cascading aborts. To ensure the serializability of transactions, we force the recoverability relationship between transactions to be acyclic.