In this paper, we propose a spatio-temporal data aggregation protocol in Vehicular Delay Tolerant Network (VDTN). We focus on Asynchronous Vehicular Crowdsensing Service (AVCS) to collect volume sensor data (e.g., images captured by on-board cameras) from VDTN-enabled vehicles. In AVCS, it is critical to cope with the huge redundant traffic generated by a large number of vehicles. We propose a novel protocol to aggregate volume spatio-temporal sensor data in Hybrid DTN data collection architecture. By assigning spatio-temporal identifiers (STI) to the aggregation targets in AVCS and extending the message exchange protocol to treat STI in VDTN, the redundant traffic can be significantly improved. Simulation results using a real taxi trace dataset showed the effectiveness of the proposed data aggregation protocol. The coverage of the crowdsensing was improved around 20-35% with 80% traffic reduction compared with the baseline aggregation protocol.

In this paper, we propose a new centralized Vehicle Delay/Disruption Tolerant Network (VDTN) architecture for vehicle volume data collections in the rural areas called "Hybrid DTN". Hybrid DTN applies Software Defined Network (SDN) into VDTN. A centralized controller controls data deliveries by the wide-area cellular network as a control-plane network and collects large volume sensor data from the vehicles to the cloud using VDTN as a data-plane network. We defined an oracle model called Delivery Status Oracle (DSO) for Hybrid DTN, in which the controller manages the delivery status of the distinct messages. By using DSO, Hybrid DTN can achieve the high data delivery success rate even for large volume sensor data collection in the rural cellular network. Simulation results using a real taxi trace dataset showed the practicalness of the Hybrid DTN with DSO; a combination with Epidemic routing algorithm hugely improved delivery completion rate with smaller delivery latency and smaller traffic.

We present the LASK protocol and its platform implementation that supports distributed k-Nearest Service Discovery. LASK achieves scalable and locality-aware name-based service discovery and routing for the target nodes avoiding redundant lookup message exchanges across the edge networks.

Edge computing enables application services to leverage low-latency responses. This paper describes the design of infrastructure-as-a-service for edge computing (EC-IaaS) to realize multiple application services on the same edge servers. In the provision of application services, it is important for application service providers (ASPs) to satisfy proximity constraints regarding the edge servers hosting virtual machines (VMs). State-of-the-art methods to provide VMs satisfying these proximity constraints require ASPs to have a large amount of information from an EC-IaaS provider, which incurs large computational and communication overheads. Otherwise, the edge servers will incur the large power draw. In this paper, we propose a virtual region model. The virtual region model abstracts the details of an edge computing infrastructure. We extend the VM request and placement processes based on the virtual region model, to support dynamics of end-user devices. We confirm that the amount of information, which ASPs receive from an EC-IaaS provider, is the same level of the least amount required by the state-of-the-art methods. Additionally, the power draw of edge servers is the same as the minimum in the state-of-the-art methods, and this level is even sustained when the number of end-user devices changes in a city.

This paper presents an overlay network protocol called “LASK: Locality-Aware Service discovery protocol for K-nearest search”, for supporting scalable and locality-aware distributed k-Nearest Service Discovery (kNSD). The kNSD provides a lookup function to find k service nodes located close to the requester. Importance of such function increases under the situation of growth in size and the diversity of the resources in the so-called edge computing environments. In LASK, efficient cooperative routing algorithm is implemented utilizing the two-dimensional structure of key-order preserving structured overlay networks that are constructed according to the network topology and the name of the service. This structure achieves scalable and locality-aware routing of the lookup messages to the matched nodes avoiding redundant data transmissions across the edge networks. Extensive simulation evaluations show that LASK could achieve small discovery latency in the large-scale edge computing environments; a typical kNSD took less than 5 ms in a data center network model and less than 100 ms in the Internet model even when there are 100,000 nodes in the entire network.

In this paper, we propose a novel dynamic data flow platform for Internet of Things (IoT) applications in edge computing environments. To avoid the overloads on network and computational resources that are caused by IoT applications, the proposed platform replicates processes and changes the structure of the data flow dynamically on the distributed computational resources located at network edges and data centers. The proposed platform adds the notion of "index" to the existing distributed topic-based pub/sub (TBPS) messaging method to support flexible data flow definitions and process allocations. In addition, we propose a peer-to-peer-based data stream routing algorithm called "Locality-Aware Stream Routing (LASR)" which can change the data stream destination dynamically on extended TBPS in accordance with structural changes of a data flow, considering network communication localities. By simulation, we confirm that our proposed platform using LASR can change the data stream destination quicker than existing methods with a small overhead in an edge computing environment, even when running applications involving video analysis data flow with a high data rate.

Edge computing enables application services to leverage low-latency responses, e.g., for augmented reality. Infrastructure-as-a-service for edge computing (EC-IaaS) allows multiple application service providers (ASPs) to use a virtualized resource (i.e., virtual machines (VMs) and virtual networks) for various applications services on the common edge computing infrastructure. For ASPs, the proximity of the VMs to the end-user devices (i.e., the location of the host edge servers) is a significant factor in guaranteeing low-latency response times. The existing method allows ASPs to select the edge server locations so that the VMs satisfy the proximity requirement. However, the amount of information transmitted from the EC-IaaS provider to the ASP becomes large, because it has to include the latencies between all edge server locations and all wireless base stations. We propose a method that reduces the amount of information while ensuring that ASPs obtain VMs that satisfy the proximity requirement. An EC-IaaS provider suggests groups of wireless base stations that have VMs within a suitable proximity, which helps the ASPs to estimate the necessary number of VMs. Instead of the ASP, the EC-IaaS provider determines which edge servers should contain the VMs. We confirm that the proposed method reduces the transmitted information. The EC-IaaS provider also has a benefit that detail latency information is kept closed to ASPs and the power draw is sustained compared to the existing method.

Summary The Pacific Rim Application and Grid Middleware Assembly (PRAGMA) is an international community of researchers that actively collaborate to address problems and challenges of common interest in eScience. The PRAGMA Experimental Network Testbed (PRAGMA‐ENT) was established with the goal of constructing an international software‐defined network (SDN) testbed to offer the necessary networking support to the PRAGMA cyberinfrastructure. PRAGMA‐ENT is isolated, and PRAGMA researchers have complete freedom to access network resources to develop, experiment, and evaluate new ideas without the concerns of interfering with production networks. In the first phase, PRAGMA‐ENT focused on establishing an international L2 backbone. With support from the Florida Lambda Rail, Internet2, PacificWave, Japan Gigabit Network, and TaiWan Advanced Research and Education Network, PRAGMA‐ENT backbone connects openflow‐enabled switches at University of Florida, University of California, San Diego, Nara Institute of Science and Technology (Japan), Osaka University (Japan), National Institute of Advanced Industrial Science and Technology (Japan), and National Applied Research Laboratories (Taiwan). The second phase of PRAGMA‐ENT consisted of an evaluation of technologies for the control plane that enables multiple experiments (ie, OpenFlow controllers) to coexist. Preliminary experiments with FlowVisor revealed some limitations leading to the development of a new approach, called AutoVFlow. This paper describes our experience in the establishment of PRAGMA‐ENT backbone (with international L2 links), its current status, and plans for the control plane. Discussion of preliminary application ideas, including optimization of routing control; multipath routing control; extending the backbone using overlay network; and remote visualization are also discussed.

A dynamic circuit network (DCN) is a production advance bandwidth reservation service. The majority of DCN operators currently provide a single class of service (CoS), e.g., either a guarantee of bandwidth or no guarantee of quality of service. Although single CoS provisioning is viable and expedient, multiple classes of service (multi-CoS) provisioning offers greatly superior practice, including quality-based charging for a commercial service and efficient resource management. This paper proposes a multi-CoS architecture with bandwidth and delay guarantees in a DCN, more specifically, the deployment of On-Demand Secure Circuits and Advance Reservation System (OSCARS) over multiprotocol label switching (MPLS) networks. The main contribution of this paper is a guaranteed bandwidth-and-delay class where the system can compute the path which can satisfy the given bandwidth and delay constraints. Our system was tested on practical routers.

Bandwidth and latency are two major factors that contribute the most to network application performance. Between each pair of switches in a network, there may be multiple paths connecting them. Each path has different properties because of multiple factors. Traditional shortest-path routing does not take this knowledge into consideration and may result in sub-optimal performance of applications and underutilization of network. We proposed a concept of "bandwidth and latency aware routing". The idea is that we could improve overall performance of the network by separating application into bandwidth-oriented and latency-oriented application and allocate different route for each type of application accordingly. We also proposed a design of this network system implemented using Open Flow. Routes are calculated from monitored information using Dijkstra algorithm and its variation. To support our design, we show a use case in which our design performs better than traditional routing as well as evaluation results.

In the era of cloud computing, data centers that accommodate a series of user-requested jobs with a diversity of resource usage pattern need to have the capability of efficiently distributing resources to each user job, based on individual resource usage patterns. In particular, for high-performance computing as a cloud service which allows many users to benefit from a large-scale computing system, a new framework for resource management that treats not only the CPU resources, but also the network resources in the data center is essential. In this paper, an SDN-enhanced JMS that efficiently handles both network and CPU resources and as a result accelerates the execution time of user jobs is introduced as a building block technology for such a HPC cloud. Our evaluation shows that the SDN-enhanced JMS efficiently leverages the fat-tree interconnect of cluster systems running behind the cloud to suppress the collision of communications generated by different jobs.

The approach of deploying hybrid wide-area Open-Flow networks is essentially important for gradual deployment of the OpenFlow technology into existing wide-area networks because it is impractical to replace such wide-area networks with OpenFlow-enabled ones at once. On the other hand, the design, deployment, and operation of such hybrid OpenFlow networks are often conducted intuitively without in-depth technical considerations. In this paper, we discuss systematically the technical and architectural aspects of hybrid OpenFlow networks based on the experiences obtained so far in the development of our wide-area hybrid OpenFlow networks on JGN2plus/JGN-X, which is nation-wide network testbed in Japan. We also describe the design and operation of RISE (Research Infrastructure for large-Scale network Experiments) on JGN-X, whose objective is to support a variety of OpenFlow network experiments.

Virtual Machine Clusters tend to increase the amount of unused memory. We propose a remote swap management framework in a VM cluster which configures a remote swap dynamically to running VMs according to memory usage. We explain the functional requirements and the design of the framework, and demonstrate the effectiveness of the framework using prototype implementation.

In cloud computing environment, virtual machine technology is used as a means of flexibly assigning workloads to real machines based on the workloads' static profiles. Though this technology aims at flexible and dynamic resource sharing, memory is underutilized in cloud environments. We propose a virtual swap management mechanism (VSMM) that enables flexible and dynamic memory sharing via a local area network. The VSMM can virtualize swap devices seen by a guest operating system (OS), and transparently switch underlying physical devices to that OS. In this paper, we explain the VSMM architecture and provide details on our prototype implementation using the Xen open source hyper visor, open source iSCSI implementations, and logical volume management. Through various experiments, we demonstrate that VSMM can contribute to improving process performance and maximally utilizing the equipped memory in cloud environments.

OpenFlow is a promising candidate for future Internet technology. In the US, OpenFlow-based networks have been installed on several university campuses recently. However, we perceive that there is no structured approach to the operation and management of large-scale OpenFlow-based networks and there are few actual achievements of carrying business traffic on them. The focus of this study is on developing a management scheme of global OpenFlow-based networks integrated into the existing Ethernet-based networks using the tunneling technology. We deployed a nation-wide OpenFlow-based network on the NICT JGN2plus testbed, and demonstrated a high-quality real-time video streaming application to verify the performance and reliability of large-scale OpenFlow-based networks. We also describe the details of technical issues we confronted in the trial operation.

Virtual machine monitors (VMMs), including hypervisors, are a popular platform for implementing various security functionalities. However, traditional VMMs require numerous components for providing virtual hardware devices and for sharing and protecting system resources among virtual machines (VMs), enlarging the code size of and reducing the reliability of the VMMs.This paper introduces a hypervisor architecture, called parapass-through, designed to minimize the code size of hypervisors by allowing most of the I/O access from the guest operating system (OS) to pass-through the hypervisor, while the minimum access necessary to implement security functionalities is completely mediated by the hypervisor. This architecture uses device drivers of the guest OS to handle devices, thereby reducing the size of components in the hypervisor to provide virtual devices. This architecture also allows to run only single VM on it, eliminating the components for sharing and protecting system resources among VMs.We implemented a hypervisor called BitVisor and a parapass-through driver for enforcing storage encryption of ATA devices based on the parapass-through architecture. The experimental result reveals that the hypervisor and ATA driver require approximately 20 kilo lines of code (KLOC) and 1.4 KLOC respectively.

In recent years, as the data processed by governmental or commercial organizations increases, cases involving information leak have risen. It is difficult to control information on many distributed end-point computers using conventional security mechanisms. Therefore, we have been proposed a novel secure VMM (Virtual Machine Monitor) architecture which is used as a foundation of security policy enforcement on distributed computers. This paper especially introduces Role-based Access Control (RBAC) to theID management framework in a secure VMM system. Our proposal will reduce costs for distributed policies updates. Proposed RBAC mechanism employs attribute certificates (ACs) to handle user’s roles. This paper shows design and prototype implementation based on PKI-based ID card and proven open source VMM software, QEMU.

We are developing a system framework for an Internet con-tent distribution service synchronized with TV broadcasting programs. To achieve high service scalability, we propose a service system model that divides data transmissions into real-time ones and the others. Real-time control data are delivered to the clients in a push style; there are no request messages sent by the clients to the server to achieve high communication efficiency. The other data (content data) are delivered to the clients in a conventional pull style using HTTP. To avoid synchronized rushes of an enormous num-ber of HTTP requests at the server, each client shifts the download start time of a content set randomly under a cer-tain restriction depending on the type of the program, the estimated download time, the content display time, etc. 1.

It is vital to achieve a disaster recovery system that allows a backup site to take over a primary site's IT services while the primary site is down. We propose a flexible replication mechanism based on service requirements such as system performance, recovery time objective (RTO), and recovery point objective (RPO). For high flexibility, the mechanism controls the replication schedule by monitoring the application's database accesses in the database connection library and matching the accesses with previously registered access patterns. In our experiments, we confirmed that the proposed mechanism outperforms other existing mechanisms, especially in situations with network delays and packet losses

This study revisits the scalability issue of polling I/O, i.e., select () and poll (). Although polling I/O was an efficient and hence popular I/O multiplexing mechanism, it is believed inadequate today to handle tens of thousands of concurrent TCP connections because the scanning cost of such a large connection list is overwhelmingly high. However, the real problem of polling I/O is not in the semantics itself but in various implementation factors such as memory allocations and copies to process a system call, pointer operations and function calls through VFS, and wait channel management to handle events. To mitigate these overheads and reinstate polling I/O as an efficient I/O multiplexing mechanism, we have developed a fast polling I/O library based on POSIX real-time signals. This library implements the full functionality of polling I/O by managing the state transition of each connection notified with realtime signals. Traditional polling I/O has a weakness especially when it polls a small number of active connections together with a huge number of idle ones. Our polling I/O library is proved to achieve high performance in that situation.

Traditionally, Wide-Wide-Web caching systems store WWW contents in filesystem. However, general purpose filesystems have a lot of unnecessary features for caching systems accompanied with data I/O. For example, typical filesystems maintain directories, inodes and free block maps (they are know as metadata). Those operations cause notable overhead. Modem WWW caching systems employ specialized filesystems for WWW caching. Furthermore, such implementations are often optimized to particular OSs and/or hardwares. Although this approach reduces overheads of filesystem, it is not portable and is expensive. This article describes an alternative approach, user level techniques for improvement of disk I/O.