Satoshi Amamiya

Ubiquitous computing with service adaptation using peer-to-peer communication framework

Mobile devices and wireless network infrastructures will be leading users to seamlessly use peer-to-peer services and ubiquitous computing by the growing of the infrastructures. In order to realize to use peer-to-peer services and ubiquitous services seamlessly, a new framework that enables users to use peer-to-peer services and ubiquitous computing is required. Hence, this paper describes a new ubiquitous computing framework, called VPC on KODAMA, using a peer-to-peer mechanism. Virtual Private Community (VPC) is an execution environment for peer-to-peer services, and provides a framework for definition of peer-to-peer services. Peer-to-peer services in VPC are defined as policy packages that have necessary elements to provide the services. Peer-to-peer services are offered in communities by collaboration among roles that are assigned to users. KODMA provides a network infrastructure for agents. Agents in KODAMA have their own community, and represent the communities. Communities have a hierarchy structure by agents residing in other agents communities. Agents have message filtering policy, and refuse messages that are against the policy. By unifying VPC and KODAMA, a new framework that enables services to define roles and their behavior and to manage logical relationship among communities is provided. VPC on KODAMA enables users to use peer-to-peer services and appliances seamlessly with their mobile devices.

Fuce: the continuation-based multithreading processor

Current trends of research on multithreading processors are the chip multithreading (CMT), which aims to exploit thread level parallelism (TLP) and to improve performance of software built onalltraditional threading components, e.g. pthreads. However, CMT is principally a straight forward extension of conventionalall symmetric multiprocessor (SMP) techniques, and it will suffer from the same limits to scalable multithreaded processing ifallit is built only on the traditional sequential-computation-based framework. Consideringallthese limitations of sequential-processor-basedallmultithreading, we are taking another approach to developing a multithreading processor dedicated to thread level parallelism(TLP). Our processor, named Fuce, is based on continuation-based multithreading. A thread is defined as a block of sequentially ordered instructions which areall executed exclusively. Every execution of a thread is triggered by one or more events, each of which is called continuation. The hardware cost and performance of the Fuce processor areallevaluated by means of a hardware implementation on FPGA and software simulation.

A semi-structured overlay network for large-scale peer-to-peer systems

Peer-to-peer (P2P) communication and computing frameworks are important for constructing robust large-scale distributed systems. Overlay network systems use distributed hash-table (DHT) to provide scalable and efficient node search capabilities. However, the DHT-based method has a problem for the maintenance cost of dynamically changing large-scale-network, in which nodes are frequently joining and leaving. This paper proposes a novel technique of P2P communication path management. The proposed technique devises a robust semi-structured overlay network called Ordered Tree with Tuft (OTT for short). OTT provides not only efficient node searching, but also low-cost self-maintenance capabilities for the dynamically changing network. In this method, joining and leaving of a node are managed in O(1) with high probability. Furthermore, the proposed OTT-based technique can find and construct a path shorter than that on the normal ordered tree, by setting up bypass links between remote nodes on OTT.

OS Mechanism for Continuation-based Fine-grained Threads on Dedicated and Commodity Processors

Fine-grained multithreading based on a natural model, such as dataflow model, is promising in achieving high efficiency and high programming productivity. In this paper, we discuss operating system issues for fine-grained multithread programs. We are developing an operating system called CEFOS based on a dataflow based computation model. A program on CEFOS consists of zero-wait threads which run to completion without suspension once started. Firing control among such threads is performed in a dataflow manner along with continuation relations in the program. Target platforms include fuce processor, which is dedicated to fine-grained multithreading, and commodity processors such as Intel x86. In this paper, after introducing our basic model and our operating system model, we discuss implementation issues on fuce and commodity platforms. The evaluation results indicate that our approach on commodity platforms is effective in reducing overheads while our approach on a special architecture naturally exploit parallelism even in I/O handling.

A multi-thread processor architecture based on the continuation model

We are developing the Fuce processor based on the dataflow computing model. Fuce means fusion of communication and execution. In order to execute many threads with multiple thread execution units efficiently, the Fuce processor executes multiple threads using the exclusive multi-thread execution model. The core concept of the exclusive multi-thread execution model is continuation based multi-thread execution, which is derived from dataflow computing. The Fuce processor aims to fuse the intra-processor execution and inter-processor communication. The Fuce processor unifies processing inside the processor and communication with processors outside as events, and executes the event as a thread. In this paper, we introduce the architecture of the Fuce processor and evaluate the concurrency performance of a Fuce processor which we described in VHDL. As a result, we understood that the processor has concurrency capability when there is sufficient thread level parallelism.

Kodama: towards a distributed web searching

Web retrieval is now one of the most important issues in computer science, and we believe that applying multi-agent systems to this area is a promising approach. It is a straightforward idea to incorporate the idea of user modeling with machine learning methods into Web search services. Kodama, which is being developed and in use at Kyushu University, is a multi-agent-based Information Retrieval (IR) system that lets users retrieve more relevant distributed information from the Web. We report methods to agentify the Web, and to exploit User Preferences (UP) by the User Interface Agent (UIA), Interpretation Policy (IP) by the Web Page Agent (WPA) and the known Server Agents (SA) table by the SA adaptively.

Pinpoint Web Searching and User Modeling on the Collaborative Kodama Agents

The primary application domain of Kodama1 is the WWW and its purpose in this application is to assist users to find desired information. Three different categories of Kodamas agents are introduced here, Web Page Agents, Server Agents, and User Interface Agents. Kodama agents learn and adapt to the Users Preferences (UP), which may change over time. At the same time, they explore these preferences to get any relevancy with the future queries. These communities of Kodama agents autonomously achieve and update their Interpretation Policies (IP) & UP and cooperate with other agents to retrieve distributed relevant information on the Web. This paper studies ways to model users interests and shown how these models can be deployed for more effective information retrieval. In terms of adaptation speed, the proposed methods make Kodama system acts as a pinpoint information retrieval system, converges to the users interests and adapts to the users sudden change of interests.

A new approach of the collaborative user interface agents

Next generation of information systems will rely on cooperative intelligent agents for playing a fundamental role in actively searching and finding relevant information on behalf of their users in complex and open environments, such as the Internet. User interface agents (UIA) are semi-intelligent systems, which help the users to access, manage, share and exchange information. Recently, various researchers have proposed a learning approach towards building such agents and some working prototypes have been demonstrated. Such agents learn by watching over the shoulder of the user and detect patterns and regularities in the users behavior. We present a new approach of the collaborative UIA that helps the user to retrieve information that is consistent to the users need. The model provides tools and utilities for the user to manage his/her information repositories with dynamic organization and adaptation views. In order to investigate the performance of the UIA, we carried out several experiments. Through the experiments, the results ensure that the techniques of personalization, clustering the users preferences, and making use of the preferences promise to achieve more relevant information to the users queries.

An information notification model with VPC on KODAMA in an ubiquitous computing environment, and its experiment

The notification of useful information to users is necessary to keep them aware of their environment and is of particular benefit in a ubiquitous computing environment. However, the amount and nature of information to be provided depends on circumstances. Too much information will confuse users. Systems used to disseminate information in a ubiquitous computing environment should not provide all information to all users, and should instead modify information disseminated depending on circumstances and context. In addition, systems designed for use in a ubiquitous computing environment need a security function for various reasons including privacy protection. For these reasons, we have developed an information dissemination model, called VPC on KODAMA. The model enables systems to provide an appropriate amount of information to users depending on circumstances and contains a security function. Using this model, we performed a large-scale experiment involving approximately one thousand participants. This paper contains the results of and discussions regarding this experiment.

Testing of Multi-agent-based System in Ubiquitous Computing Environment

Agent technology is ready to shift from the stage in which agent architectures are proposed to the stage in which applications working in the real world are developed. Accordingly, we have developed an agent-based application integrating VPC and KODAMA, as an example of a system where agents can behave effectively in the real world. VPC on KODAMA is a framework for actualizing a ubiquitous computing environment. VPC on KODAMA has a mechanism that assigns services to user agents according to their profiles. Using VPC on KODAMA, we developed an information notification system. This system sends advertisement e-mail tailored to user profiles, including user location, to the user’s cellular phone. We have conducted a large-scale experiment in shopping malls. The results of this experiment show that agent technology is effective for applications in ubiquitous computing environments.