Tatsuya Hagino

Effects and Performance of Content Negotiation Based on CC/PP

We describe our investigation of the effectiveness ofWeb content negotiation using CC/PP. CC/PP is a proposed specification on a user-side content negotiation framework to comply with various clients. CC/PP is based on common technologies such as XML and HTTP extension and is expected to provide a generic content negotiation solution, but neither its performance nor effectiveness has ever been shown. We have implemented a CC/PP capable Web browser and a CC/PP proxy, and measured the performance of CC/PP and various different settings. The result shows that the use of indirect reference and abbreviating profile-diffs have good improvements on retrieval time when the client connection is as narrow as cellular phones. When the connection is faster than that, the use of inline encoding would be better provided that the profile size is less than a certain threshold, in our tests, approximately one-tenth the value of effective bandwidth of the client connection. The observed traffic with content conversion by CC/PP was smaller than the case without CC/PP for all tested environments, and the retrieval time is better or comparable for cellular phone clients. The result confirms that CC/PP is an effective solution for general content negotiation.

Towards Efficient Sport Data Integration through Semantic Annotation

In news genre, the sport domain is one of great interest to audiences on many occasions. The explosion of Internet these days leads to many obstacles for users in searching information due to enormous amount of data collected from multiple media streams. In this case, Semantic Web is believed to bring a solution for information integration. In this paper, we introduce a novel approach for an integration system of sports information using semantic web. Semantic annotation of sport news on the web is a way to bring success to our system. We propose an algorithm to generate semantic annotations of sport news that relies on the ontology. Named entities recognition is improved by integrating the KIM information extraction system with this ontology, which is dedicated for sport domain. The concepts and relations of ontology detected in the text are combined with these entities to capture the semantic. Initial experiments on soccer news demonstrated promising results.

Ad-hoc Filesystem: A Novel Network Filesystem for Ad-hoc Wireless Networks

In this paper, we propose a new filesystem named Ad-hoc Filesystem which targets ad-hoc wireless networks. Ad-hoc Filesystem is a serverless filesystem which automatically generates temporary shared space among multiple mobile machines when they gather in a communicable range. The generated space is kept stable as long as possible even if some machines join or leave the communicable range. We have designed an initial prototype of Ad-hoc Filesystem and have conducted preliminary evaluations on the simulator. Our design is strongly based on an assumption such that people who wish to work together would form a stable group with a single mobility behavior. Based on the assumption, our prototype distributes files among multiple machines and duplicates each on two machines, then it keeps the pairs of such machines as long as possible. We also introduced another strategy to increase file availability by exploiting each machines client cache. The simulation results show that controlling packets used in our prototype to maintain system state hardly affects the overall system performance. Also the result indicates that our strategy that keeps two replicas for each data and exploits client caching is effective to achieve high availability.

The ToM (thread on modules) microkernel

The authors have been developing a distributed operating system called Threads on Modules (ToM). It supports a programming model which is suited for distributed environments. For this purpose, modules, threads, houses, users, RPCs and capabilities are introduced. To make the ToM system as portable as possible, it was implemented using the microkernel technology. The ToM microkernel implements the minimal functionalities and the others are implemented as user level programs. Although the ToM microkernel is designed to support the ToM programming model, it has sufficient functionalities to implement other operating systems. This paper presents some of the main features of the ToM microkernel. A UNIX emulator also is presented.

Commons: resource sharing and protection in mobile appliance OS

In this paper, we propose a new operating system abstraction, called a common, for mobile appliances. We can use a common to prevent the excessive use of computational resources on appliances.Mobile appliances will have networked servers such as a http server. We need a new resource abstraction for mobile appliance OS to protect relatively precious and limited resources against service requests from the Internet.The common abstraction can restrict accesses of machine resources and maintain stability on appliances against unknown users on the Internet. We describe two applications of the common: resource allocation for a WWW server and restriction on a bad influence of DoS attacks.In the next section, we describe system requirements for mobile appliance OS and point out its needs. In Section 3, we describe the abstraction of a common. Section 4 describe the applications of commons. In Section 5 we describe current status of our project. Section 6 discusses related work and Section 7. concludes.

Comprehensive operating system for highly parallel machine

This paper describes the design and implementation of a comprehensive operating system (COS) for highly parallel machine. This is one of the research themes in the Japanese University Highly Parallel Research Project supported by the Special Grant of the Research Funds of Ministry of Education. The target machine is called JUMP-1 originally developed in this project. Several design policies are discussed, and the user service classes are defined. COS architecture is based on the micro kernel, and it includes both Partition OS and Service OS. The development environment is also discussed briefly.<<ETX>>