Satoshi Katsuno

Stream Mining for Network Management

Network management is an important issue in maintaining the Internet as an important social infrastructure. Finding excessive consumption of network bandwidth caused by P2P mass flows is especially important. Finding Internet viruses is also an important security issue. Although stream mining techniques seem to be promising techniques to find P2P and Internet viruses, vast network flows prevent the simple application of such techniques. A mining technique which works well with extremely limited memory is required. Also it should have a real-time analysis capability. In this paper, we propose a cache based mining method to realize such a technique. By analyzing the characteristics of the proposed method with real Internet backbone flow data, we show the advantages of the proposed method, i.e. less memory consumption while realizing real-time analysis capability. We also show the fact that we can use the proposed method to find mass flow information from Internet backbone flow data.

On flow distribution over multiple paths based on traffic characteristics

In traffic engineering to effectively distribute traffic flows over multiple network paths, taking into account traffic characteristics for individual flows is vital in appropriately assigning flows to network paths for achieving better delay performance of the total traffic All flows assigned to a path, in which some flows are highly bursty, equally experience a significant queuing delay when the path is highly utilized We analyze the mean, variance, and 99.5th percentile of queuing delay when two types of flows are multiplexed into a single path using an infinite buffer model with on-off state fluid flows This is carried out to find a better traffic distribution strategy over multiple paths so as to lower queuing delay incurred to multiplexed flows with distinct traffic characteristics From the analytical results, we found that trivial strategies, such as dividing individual types of flows in a proportional manner to the bandwidth of each path, or segregating distinct types of flows as much as possible, do not always achieve a good delay performance, e.g., in terms of max-min fairness Thus, the flows should be distributed considering their traffic characteristics, the number of flows, and the path bandwidths.

Traffic Characteristics-Based Flow Assignment Method for Reducing Queuing Delay

In traffic engineering, for effectively distributing traffic flows over multiple network paths, it is vital to take traffic characteristics of individual flows into account in appropriately assigning the flows to the network paths to achieve better delay performance as a whole. We have developed a traffic characteristic-based flow assignment method that reduces queuing delay in a buffer at the beginning of each network path. By taking the traffic characteristics of individual flows into account the method can effectively assigns flows to the paths so as to minimize overall queuing delay. We consider the simple case in which two types of flows with distinct traffic characteristics (buStiness) are assigned to two bandwidth-guaranteed paths. The flow assignment of our method reduces the queuing delay (at the edge router) of the worse path, by attempting to optimize the min-max fairness between the paths in terms of queuing delay. Numerical simulation showed that the method assigned the flows such that the queuing delay was reduced up to 40% compared with that when conventional path utilization based flow assignment was used. This means that taking the traffic characteristics of flows into account significantly improves the queuing delay performance when the flows have distinct characteristics.

Performance Monitoring of VoIP Flows for Large Network Operations

The monitoring of performance in VoIP traffic has become vital because users generally expect VoIP service quality that is as high as that of PSTN services. A lightweight method of processing by extracting VoIP flows from Internet traffics is proposed in this paper. Estimating delay variations and the packet loss ratio using knowledge about specific features and the characteristics of VoIP flows, i.e., the inter-packet gap (IPG) which is constant in VoIP flows, is also proposed. Simulation with actual traffic trace is used to evaluate the method, and this revealed that delay variations (IPG variance) can be accurately estimated by monitoring only a few percentage of all flows. The proposed method can be used as a first-alert tool to monitor large amounts of flows to detect signs of degradation in VoIP flows. The method can be used by ISPs to estimate whether VoIP flow performance is adequate within their networks and at ingress from other ISPs.

Measurement and analysis of multimedia application and IPv6 ADSL Internet access network

This paper presents measurements and analysis of broadband Internet access service networks using a highspeed IP meter with a GPS (Global Positioning System) timestamp component. The authors performed measurements of traffic in a commercial IPv6 access service on ADSL (asymmetrical digital subscriber line) as a typical broadband access service network. This paper describes the measurement configuration using the high-speed IP meter and reports asymmetrical characteristics of packet delay on ADSL access networks. It also presents the analysis results of multimedia applications in the Internet, and discusses the quality of service on Internet access service networks.

Distributed Flow Monitoring Tool Using Network Processor

There is an emerging requirement for real-time flow-based traffic monitoring, which is vital for quality of service (QoS) monitoring, anomaly monitoring, and traffic engineering in Internet service provider networks. In this report, we discuss a distributed architecture for real-time and scalable flow monitoring tools and implement a prototype tool using network processor (NP). The contributions of the present paper are two-folds. First, we clarify the elemental functions that are required and commonly used by passive flow monitoring tools, and refine a distribution-processing architecture for combining these functions based on our previous work in order to cope with heavy-load input traffic. Second, we implement the packet distribution component, which plays a core role and might be bottleneck in the proposed architecture, by using pipeline-mode NP hardware, and experimentally prove its applicability to 10 Gbit/s-class input traffic

An implementation and evaluation of IPv6 end-to-end secure communication system for closed members

Establishing secure channels between a pair of hosts is an attractive way when the direct exchange of important data is required. IPsec provides various security services for traffic at the IP layer so we can use it to establish secure channels between a pair of hosts. However, IPsec is difficult to use because there are many parameters to set up for secure channels, and the configuration is complicated, hence our previous proposal of an automatic configuration method to set up the end-to-end secure channels between a pair of hosts; facilitating the use of IPsec. In this paper, we overview our proposed method and describe its implementation and evaluation. In addition, we show how the IPsec setup for 10 hosts on a host implementing our method is completed within a few seconds or so

On the influence of network characteristics on application performance in the grid environment

In the Grid computing, it is a key issue how limited network resources are effectively shared by communications of various applications in order to improve the application-level performance, e.g., to reduce the completion time of each application and/or a set of applications. In fact, the communication of an application changes the condition of network resources, which may, in turn, affect the communications in other applications, and thus may deteriorate their performance. In this paper, we examine the characteristics of traffic generated by some typical grid applications, and how the round-trip time and the bottleneck bandwidth affect the application-level performance (i.e., completion time) of these applications. Our experiments show that the impact of network conditions on the application performance and the impact of application traffic on the network conditions are considerably different depending on the application. Those results suggest an effective network resource allocation should take network-related properties of individual applications into consideration.