Saneyasu Yamaguchi

The multi-agent system for dynamic network routing

The single-hop communication method of the current wireless network cannot meet new demands in the domain such as intelligent transport systems (ITS) and Home LAN. Even though the ad-hoc network architecture is expected to solve this problem, the nature of a dynamic topology makes this routing hard to realize. This paper introduces a new ad-hoc routing algorithm, which is inspired by the cooperating mobile agents for dynamic network routing of Miner et al. (see Software agents for future communicating systems, p.287-304, Springer, 1999). In their system, some control agents explore the network and update routing tables on their own knowledge. Using these routing tables, other agents deliver messages. They considered the feasibility of the agent-based routing system, but did not refer to an efficient algorithm. We consider that algorithm without increasing the network load. We propose multiplying each entry in the routing table to store much more information from agents and evaluating them to make better use of information, which succeeded in raising the network connectivity by about 40% by simulation.

Trace system of iSCSI storage access

In this paper, an IP-SAN access trace method is proposed and its implementation is presented. IP-SAN and iSCSI are expected to remedy problems of fibre channel (FC)-based SAN. Because servers and storage cooperatively work with communications through TCP/IP layer in IP-SAN system, an integrated analysis of both sides is considered to be significant for achieving better performance. Our proposed system can precisely point out the cause of performance degradation when IP-SAN is used for a remote storage access. In the experiment of parallel iSCSI access in a long-delayed network, the total performance is limited by a parameter in an implementation of the SCSI layer in the iSCSI protocol stack. Based on the result obtained with our IP-SAN access trace system, the parameter in the layer is modified. As a result, four times performance improvement is achieved compared with the default value case. Thus it is effective to monitor all the layers in the iSCSI protocol stack and execute an integrated analysis, using our system.

Autonomous load balance system for distributed servers using active objects

We have implemented an autonomous load balancing system for distributed servers using the active object model. This system consists of distributed servers. Each server has a load balancer which monitors the server load, and controls the load, communicating with other load balancers. Distributed load balancers communicate in a peer-to-peer way, and a client/server model is not suitable, the active object model is very suitable. Therefore we have implemented a Java library CAPE for implementing peer-to-peer based distributed active object systems. Using the CAPE library, the load balancing system can be implemented nicely.

Trace system of iSCSI storage access and performance improvement

In this paper, an IP-SAN access trace method is proposed and its evaluation is presented. IP-SAN and iSCSI are expected to remedy problems of Fibre Channel (FC)-based SAN. Servers and storage cooperatively work with communications through TCP/IP in IP-SAN system, thus an integrated analysis of both sides is considered to be significant for achieving better performance. n nOur system can precisely point out the cause of performance degradation when IP-SAN is used for a remote storage access. In experiment of parallel iSCSI access in a high-latency network, the total performance is limited by a parameter in an implementation of the SCSI layer in the iSCSI protocol stack. Based on the result obtained with our IP-SAN access trace system, the parameter in the layer is modified. As a result, more than 30 times performance improvement is achieved compared with the default value case. Thus it is effective to monitor all the layers in the iSCSI protocol stack and execute an integrated analysis, using our system.

Performance Analysis of iSCSI Middleware Optimized for Encryption Processing in a Long-Latency Environment

Since IP-SAN allows us to reduce initial costs and management costs of storage by using IP networks, iSCSI, common protocol in IP-SAN, is becoming more important. Although one of the key issues for iSCSI is a security measure to access remote storage over IP networks, this is not necessarily established yet. iSCSI can employ IPsec that offers a function of strong encryption. However, the IPsec encryption processing degrades throughput of storage access and increases its CPU load. In addition, it is difficult to execute encryption processing efficiently, because IPsec layer is located in a lower-level. In this paper, we implemented a middleware system of executing encryption processing in the upper-layer for the communication using iSCSI securely and effectively. Moreover, we simulated the idea of optimization of encryption processing and evaluated its performance by running parallel processes in a long-latency network. As a result, optimization method of encryption processing is more efficient than that of IPsec in a long-latency network

TCP Congestion Window Controlling Algorithms on iSCSI Sequential Read Access

iSCSI is one of the most promising technologies in IPSAN, which decreases a storage management cost. Congestion Window, one of TCP parameters, has a close relationship with system performance in accessing remote storage with iSCSI. In this paper, we propose a dynamic CongestionWindow control method to stabilize the uneven behavior of throughput observed on the iSCSI storage access, and examine the algorithms to make the Congestion Window stabilized as soon as possible in the dynamic Congestion Window control method. We have adopted a linear search method, binary search method, and Congestion Window notification method. These algorithms are implemented on the experimental system and evaluated.

Performance Evaluation of iSCSI System Optimized for Encryption Processing in the Upper Layer

iSCSI protocol, used in building IP-based storage networks, is becoming more important because it realizes consolidation of storage at low cost. Security is a critical issue for the iSCSI protocol, on which remote storage is accessed over the IP networks. iSCSI can employ IPsec that offers a function of strong encryption. However, IPsec encryption processing degrades the performance of storage access and increases the CPU load of the server. In this paper, to perform secure storage access on iSCSI networks, we propose the storage access method of encrypting in the upper layer and its optimization, instead of using IPsec encryption that leads to the performance degradation. We implemented our proposed system, and experimentally evaluated our system accessed with multi process for simulating optimization of encryption pre-processing. As a result, our proposal, optimizing of encryption pre-processing is more efficient than the method using IPsec.

iSCSI analysis system and performance improvement of iSCSI sequential access in high latency networks

IP-SAN and iSCSI are expected to remedy the problems of FC-based SAN. iSCSI has a structure of multilayer protocols. A typical configuration of the protocols to realize this system is as follows: SCSI over iSCSI over TCP/IP over Ethernet. Thus, in order to improve the performance of the system, it is necessary to precisely analyze the complicated behavior of each layer. In this paper, we present an IP-SAN analysis tool that monitors each of these layers from different viewpoints. By using this analysis tool, we experimentally demonstrate that the performance of iSCSI storage access can be significantly improved by more than 60 times.

Cooperative background task spaces and its evaluation

Recently, not only servers computer but also computers for the individual is connected to the Internet. These computers are not always loaded. Therefore, there are many idle computer resources. We propose a system using computer resources only while they are idle. In other words, a task is given to a computer when its user leaves it. The task is processed while the computer is left idle. And the task is moved to another idle computers when the user is back and resumes the primary works. The system can process tasks using idle computer resources without hindering primary work. The system has functions of automatic load balancing, dynamic task migration, etc. We did experiments and evaluated these functions. As a result, it was confirmed that these functions worked suitably.