Data-centric networking has recently been getting increased attention. A representative design of data-centric networking is CCN (Content-Centric Networking), which routes packets within a network based on their content identifiers. CCN is basically designed to be open because ease of data reuse is one of the greatest advantages of data-centric networking. However, being used for real-world networking, completely open data-centric networking is not sufficient. It is required to realize closed communication within a group of users. In this paper, we propose Virtual Content-Centric Networking (VCCN), which realizes closed communication within a group of users with CCN router virtualization. This paper presents four building blocks of VCCN: extension of the content identifier, CCN router virtualization, packet transport between virtualized CCN routers, and Social Network Services cooperative user/group identification. Moreover, we implemented VCCN's basic features by extending the CCNx software and performed a preliminary performance evaluation of our VCCN implementation.

Gradient-based routing, where each node calculates a metric that indicates how useful a node might be in relaying messages to a sink node and transmits messages according to the metric, is one of promising approaches for Delay Tolerant Mobile Sensor Networks. However, existing gradient-based routing methods do not consider node mobility to form their gradient and this may result in inefficient message relays and degradation in their performance. In this paper, we discuss how node mobility affects message delivery in gradient-based routing and propose a gradient-based routing method that incorporates node mobility into its gradient to reduce the effect of inefficient message relays. The key idea of our proposal is to distinguish nodes leaving from a sink node from nodes approaching to a sink node. Since those leaving nodes are less useful to relay messages to a sink node, our proposed method prevents nodes from transmitting messages to nodes leaving from a sink node. Through simulations, we show that our proposal decreases the average message delivery delay under various node mobility models. Moreover, our proposal reduces the average message delivery delay by up to 50% in the case that nodes move straightly.

Virtual machine live migration, which migrates a virtual machine between data centers, is studied as a way to improve quality of services hosted on clouds. Meanwhile, traffic engineering is performed in networks that connect geographically-dispersed data centers. These two controls are originally designed and operated individually. Though it is naturally expected that integrating virtual machine live migration and the traffic engineering could result in a good overall performance, the effectiveness of such an integrated control has not been well understood. In this paper, we therefore quantitatively investigate its effectiveness. We first formulate an integrated control and an individual control as mixed integer programming problems in which the objective function is minimization of the average link delay in the network. Through numerical examples, we show that the integrated control can reduce the average link delay by at most 24 % and it can accommodate as 1.3 times much as incoming traffic compared with the individual control.

In this paper, we present our vision on large-scale and dynamic social network analysis in real environment, which is expected to be enabled by introduction of large-scale heterogeneous sensors in ambient environment. We address challenges toward realization of large-scale dynamic social network analysis in real environment, and discuss several promising applications. We finally present our preliminary experimental results of dynamic social network analysis for six-person social gatherings in real environment.

In this paper, we propose a novel communication paradigm called group-oriented communication. Different from conventional unicast-based communications, group-oriented communication is entirely based on group-based communication. Our group-oriented communication is essentially a type of many-to- many communication, but it realizes any type of communications including one-to-one, one-to-many, many-to-one and many-to- many communications based on group-based communication. With our group-oriented communication, diverse social activities can be shifted into a communication network in a straightforward way, and users' requirements on security/reliability can be fulfilled. In this paper, we first qualitatively discuss advantages of our group-oriented communication by comparing with the conventional IP-based network. We then discuss four design goals of a network architecture for our group-oriented communication: supporting dynamic entity/group, supporting address operation expression, realization of entity/group find ability, and realization of security. After carefully examining these design goals, we design a network architecture for realizing our group-oriented communication. Through quantitative evaluations, we show that the network architecture for our group-oriented communication should be packet-based, that reachability control is the core networking technology, and that the network architecture should have the two-layer structure consisting of transport and control layers.

It has been recognized that in the current Internet it is important to control not only congestion caused by overload of the data plane but also congestion caused by overload of the control plane. In particular, it is important to avoid congestion from retrial traffic. This paper focuses on RSVP-based communication services. In general, users reattempt to set up connections not only when transmission systems are overloaded but also when processing systems in the network are overloaded. The latter is caused by user psychology: an increase in the waiting time for the processing to be completed tends to increase his or her reattempts. Thus, it is important to manage both transmission and processing resources properly. In this paper, we introduce a model that takes account of both transmission and processing systems, and we examine the behavior of retrial traffic. In addition, we compare the system stability achieved by two different resource management methods, the hard-state method and the soft-state method.

XCP (eXplicit control protocol) has been proposed as an efficient transport protocol for a wide-area and high-speed network. XCP is a transport-layer protocol that performs congestion control using explicit feedback from routers. In the literature, many simulation-based performance studies of XCP has been performed. However, the effect of traffic dynamics on the XCP performance has not been investigated. In this paper, through simulation experiments, we first show that XCP has the following problems: (1) utilization of the bottleneck link is lowered due to XCP traffic dynamics, and (2) in environment where non-XCP traffic and XCP traffic coexist, control of XCP becomes unstable. We then propose XCP-IR (XCP with increased robustness) that operates efficiently even for dynamic traffic. Through simulation experiments, we show that XCP-IR operates efficiently even for dynamic traffic.

In this paper, we analyze stability of XCP (explicit control protocol) in a network with heterogeneous XCP flows (i.e., XCP flows with different propagation delays). Specifically, we model a network with heterogeneous XCP flows using fluid-flow approximation. We then derive the conditions that XCP control parameters should satisfy for stable XCP operation. Furthermore, through several numerical examples and simulation results, we quantitatively investigate effect of system parameters and XCP control parameters on stability of the XCP protocol. Our findings include: (1) when XCP flows are heterogeneous, XCP operates more stably than the case when XCP flows are homogeneous, (2) conversely, when variation in propagation delays of XCP flows are very large, operation of XCP becomes less stable, and (3) output link bandwidth of an XCP router is independent of stability of the XCP protocol.

In this paper, we propose a flow-level simulator called FSIM (Fluid-based SIMulator) for performance evaluation of large-scale networks, and verify its effectiveness using our FSIM implementation. The notable features of our flow-level simulator FSIM are its accuracy and fast simulation execution compared with conventional flow-level simulators. For improving simulation accuracy, our flow-level simulator FSIM utilizes accurate fluid-flow models. For accelerating simulation execution speed, our flow-level simulator FSIM adopts an adaptive numerical computation algorithm for ordinary differential equations. Another notable feature of our flow-level simulator FSIM is its compatibility with the existing network performance analysis tool. In this paper, through several experiments using our FSIM implementation, we evaluate the effectiveness of our flow-level simulator FSIM in terms of simulation speed, accuracy and memory consumption. Consequently, we show that our flow-level simulator FSIM outperforms a conventional flow-level simulator; i.e., it realizes approximately 100% faster simulation with higher accuracy and less memory consumption than a conventional flow-level simulator.

GridFTP has been used as a data transfer protocol to effectively transfer a large volume of data in grid computing. GridFTP supports a feature called parallel data transfer that improves throughput by establishing multiple TCP connections in parallel. However, for achieving high GridFTP throughput, the number of TCP connections should be optimized based on the network status. In this paper, we propose an automatic parallelism tuning mechanism called GridFTP-APT (GridFTP with automatic parallelism tuning) that adjusts the number of parallel TCP connections only using information measurable in the grid middleware. Through simulation experiments, we demonstrate that GridFTP-APT significantly improves the performance of GridFTP in various network environments.

In recent years, IP-based virtual private networks (IP-VPNs), which provide a virtual privately owned network over an IP network, have attracted attention. With existing IP-VPNs, however, there is a serious problem that fairness among IP-VPN customers is not satisfied. In our previous work, we have proposed an IP-VPN fairness control mechanism called I2VFC (Inter-and Intra-VPN Fairness Control) that realizes fairness among IP-VPN customers. In this paper, we quantitatively show effectiveness of our I2VFC using simulation experiments and prototype system experiments. Focusing on inter-VPN fairness, intra-VPN fairness, and scalability, we extensively analyze the performance of I2VFC. Consequently, we show that I2VFC can realize both inter-and intra-VPN fairness under diverse control parameter configurations, indicating robustness and parameter insensitivity of our I2VFC. We also show that I2VFC has a practically sufficient scalability in terms of the transfer speed and the number of VPNs accommodated. For instance, measurement results using our prototype system show that with a modern desktop computer, I2VFC can support approximately 1.6 [Gbit/s] bandwidth and 1,300 numbers of VPNs.

In this paper, we propose a network model partition method called QD-PART (Quasi-Dynamic network model PARTition method) for accelerating parallel network simulation. The key of QD-PART is to utilize the fact that a network simulation is typically repeated several times with the same parameter set for estimating the confidence interval of steady state measures. QD-PART gradually optimizes partition of a network model based on past simulation results such as the total simulation time, CPU usage of computing resources, and traffic intensity (i.e., the number of packets transmitted) of each link. At the end of each parallel simulation run, QD-PART re-partitions the network model based on such information aiming at minimizing communication overhead among computing resources and balancing load of sub-network models executed on computing resources. Through several experiments using a parallel-distributed network simulator, we show how parallel network simulation can be accelerated using QD-PART by gradually improving the network model partition.

In recent years, IP-based virtual private networks (IP-VPNs), which provide a virtual privately owned network over an IP network, have attracted attention. With existing IP-VPNs, however, there is a serious problem that fairness among IP-VPN customers is not satisfied. In this paper, we first discuss design objectives of a control mechanism for achieving fair IP-VPN services: achieving inter-VPN fairness, achieving intra-VPN fairness, easy deployment into existing IP networks, and achieving a high scalability. We then propose an IP-VPN fairness control called 12FVC (inter-and intra-VPN fairness control) for realizing a fair IP-VPN service in a scalable way. The core of 12VFC is an AIMD (additive increase and multiplicative decrease) window flow control operating among IP-VPN service provider's edge routers. 12VFC has the advantage that an IP-VPN service provider can arbitrarily specify inter-VPN fairness criteria by utilizing analytic results of AIMD window flow control. Moreover, 12VFC can be easily deployed into existing IP networks by simply modifying edge routers. Through several simulation experiments, we demonstrate that 12VFC realizes both inter-VPN fairness and intra-VPN fairness with extremely high accuracy.

The proposed metric for evaluating class library interfaces measures the balance between generalization and specialization. The key idea of the proposed metric is the fact that polymorphic operations satisfy Zipf's law, which is known as an approximating equation for the frequency of words in natural language. The metric shows the global property of an entire class library, independent of size (making it possible to compare class libraries of different sizes and different versions), and to avoid the problems associated with using the depth of the inheritance graph as the metric. The results of application of this metric to two upgrades of the InterViews C++ class library agree with the results of manual analysis on three points: the metric value for the groups whose interfaces are generalized are larger than the metric value of other groups which are not as generalized, the metric value increases when a class interface is generalized by creating abstract classes, and the metric value decreases when new specialized-functional extension classes are added to the class library.

This paper proposes a new problem called the dynamic Steiner tree problem. Interest in the dynamic Steiner tree problem is motivated by multipoint routing in communication networks, where the set of nodes in the connection changes over time. This problem, which has its basis in the Steiner tree problem on graphs, can be divided into two cases: one in which rearrangement of existing routes is not allowed, and a second in which rearrangement is allowed. For the nonrearrangeable version, it is shown that the worst-case performance for any algorithm is at least $\frac{1}{2}\lg n$ times the cost of an optimum solution with complete rearrangement. Here n is the maximum number of nodes to be connected. In addition, a simple, polynomial time algorithm is present that has worst-case performance within two times this bound. In the rearrangeable case, a polynomial time algorithm is presented with worst-case performance bounded by a constant times optimum.

This paper clarifies the relation between the diameter k and the edge or node connectivity Ce or c,, of digraphs. The following two inequalities are derived: where n is the number of nodes, d is the maximum degree, and d is the minimum degree.

This paper proposes a simple procedure for the design of small-diameter graphs. It can be used to construct a directed graph whose diameter is less than or equal to that of any previously proposed graph.

This paper discusses adaptive capacity allocation for high-usage groups (direct routes) in telephone networks so as to best maintain performance for each hour's traffic load. We present two algorithms for this adaptive allocation scheme. The first, called a capacity allocation algorithm, determines capacity allocation in a manner that minimizes the maximum value of point-to-point blocking probabilities for each hour's traffic load. The second, called a frame assignment algorithm, determines the channel connection pattern with a minimum of changes from the old pattern, for a new capacity allocation.

This paper proposes a simple algorithm for the graph design of small-diameter networks. For given nodes n and degree d, this algorithm can be used to construct a directed graph with diameter ⌈logd n⌉, which is at most one larger than the lower bound ⌈logd (n(d − 1) + 1)⌉ − 1. Its average distance is also close to the lower bound. The algorithm can be used to construct a directed graph with smaller diameter, compared with any other conventional methods, when n is larger.