Proceedings 22nd International Conference on Distributed Computing Systems Workshops

Describes work in progress on wearable computing configurations which provide audio and visual output based on the position and orientation of the user. We introduce the 'eSleeve'-a wearable wrist computer with position and heading sensors combined with a user interface employing speech recognition and a small display. Applications described include searching a database of locations, and implementing a minimal augmented reality system. The effectiveness of different approaches at conveying location based information is discussed and we also describe our continuing research into position based audio and visual interfaces.

A key issue when designing and implementing largescale publish/subscribe systems is how to efficiently propagate subscriptions among the brokers of the system. Brokers require this information in order to forward incoming events only to interested users, filtering out unrelated events, which can save significant overheads (particularly network bandwidth and processing time at the brokers). In this paper we contribute the notion of subscription summaries, a mechanism appropriately compacting subscription information. We develop the associated data structures and matching algorithms. The proposed mechanism can handle event/subscription schemata that are rich in terms of their attribute types and powerful in terms of the allowed operations on them. Our major results are that the proposed mechanism (i) is scalable, with the bandwidth required to propagate subscriptions increasing only slightly, even at huge-scales, and (ii) is significantly more efficient, up to orders of magnitude, depending on the scale, with respect to the bandwidth requirements for propagating subscriptions.

Emulation is an efficient method in mobile ad hoc network research, which provides a good approach to accurately evaluate the MANET routing protocol large scale. In this paper, we describe the modeling and implementation of a new low cost MANET emulator called NE based on Ethernet environment. NE provides mobile multi-hop wireless emulation including incomplete connectivity, random frame loss, bandwidth limitation and communication delay. NE is suitable for middle scale MANET emulation where the number of nodes is about one hundred.

We focus on how to generate well-written texts to describe an object from a database, and propose several strategies that are needed in generation stages. To build reliable generation rules, we performed corpus analysis through annotating descriptive texts. This paper also describes an implemented text generation system called XExplainer, which can dynamically produce a description of an object in Korean. XExplainer was applied to two domains-a home shopping database and a business administration database-to show that it can be applied to any domain as long as the information is provided in the required format. The generated texts were evaluated by humans using several criteria, such as content completeness, structural coherence, expression conciseness, and text layout.

As more software is being developed and distributed via the Internet, the number of accesses to file servers, such as FTP servers or WWW servers, is increasing. To reduce concentrated access to the original file server, mirror servers which have the same directories and files as the original server stores have been introduced. However, inconsistency among mirror servers and the original server is often observed because of transfer latency, traffic congestion on the network, and management policy of the mirror servers. Hence, we develop an intermediate FTP proxy server system PFTPD (Proxy FTP Daemon) which guarantees the freshness of files as well as dispersal of the access concentrated on the original FTP server. We found that the proposed system can disperse accesses to the mirror servers with a little overhead. This technology can assure that files are the latest version and adapt to heterogeneous network environments such as the Internet.

Bài báo giới thiệu một nguyên mẫu của hệ thống hỗ trợ chăm sóc sức khỏe đeo được mang tên 'LifeMinder', bao gồm một mô-đun cảm biến đeo được hình dạng đồng hồ đeo tay và một trợ lý số cá nhân (PDA). Mô-đun cảm biến đeo được, được trang bị các cảm biến như cảm biến gia tốc, máy đo nhịp tim, nhiệt kế, điện cực phản xạ điện galvanic (GSR) và mô-đun Bluetooth để giao tiếp với PDA, theo dõi ngữ cảnh của người dùng: tình trạng sức khỏe, chuyển động và hành vi. Hệ thống sử dụng thông tin này để hướng dẫn người dùng thực hiện chăm sóc bản thân hàng ngày theo thời gian thực. Chăm sóc chế độ ăn uống và chăm sóc vận động đặc biệt quan trọng để ngăn ngừa "bệnh tật liên quan đến lối sống". Các tác giả đã phát triển các thuật toán để nhận diện chuyển động của người dùng từ chuyển động cổ tay và phát hiện sự bắt đầu của bữa ăn từ nhịp tim và giá trị GSR. Độ chính xác của cả hai thuật toán khoảng 90%.

In this paper QoS control of a real-time multimedia communication system under heterogeneous environments of wired and wireless networks is proposed. In our suggested system, as channel coding the FEC (forward error correction) method with Reed-Solomon coding is introduced to reduce the packet error rate on the wireless network. On the other hand, as source coding, transcoding methods including transformation of various video codings such as M-JPEG, MPEG and Quicktime, control of Q-factor within a frame, frame rate and color depth are introduced to maintain the required QoS, particularly the end-to-end throughput. The increases of the required bandwidth by redundant packet addition with FEC can be suppressed by the transcoding functions while the packet error rate is reduced to the accepted value. In order to verify the functionality and the efficiency in our suggested system, numerical simulation was carried out. As a result, our suggested system by combination of transcoding and FEC could correct the packet error rate with an order of 10/sup 9/ while maintaining the frame rate and the amount of data transform constant.

This paper describes the implementations and test results of elliptic curve cryptography (ECC) and elliptic curve digital signature algorithm (ECDSA) algorithms based on Java card. A 163-bit ECC guarantees as secure as the 1024-bit Rivest-Shamir-Adleman (RSA) public key algorithm, which has been frequently used until now. According to our test results, the 163-bit ECC processing time is about five times faster compared with the 1024-bit RSA, and the amount of resource usages of ECC is smaller than RSA. Therefore, ECC is more appropriate for use on secure devices such as smart cards and wireless devices with constrained computational power consumption and memory resources.

Lecture-on-demand (LOD) multimedia presentation technologies in networks are most often used in communication services. Examples of those applications include video-on demand, interactive TV and communication tools in a distance learning system, etc. In this paper, we describe how to present different multimedia objects on a Web-based presentation system. The distributed approach is based on an extended timed Petri net model. Using characterization of extended media streaming technologies, we developed a Web-based multimedia presentation system. For a real-world example, suppose a well-known teacher is giving a lecture/presentation to his students. Because of time constraints and other commitments, many students cannot attend the presentation. The main goal of our system is to provide a feasible method to record and represent a lecture/presentation. Using a browser with windows media services allows those students to view live video of the teacher giving his speech, along with synchronized images of his presentation slides and all the annotations/comments. In our experience, this approach is sufficient for distance learning environments.

Scaling applications to large networks and an increasing number of users has been for years a technical challenge. Today, technologies are well known to scale applications to local networks but scaling to large networks with high latency is still a challenge. The DLBS (dynamic load balancing service) brings new solutions regarding large scale load balancing for middleware based applications. DLBS offers a multi-criteria and easily customizable load balancing service. It consists of a scalable monitoring infrastructure, a connection manager (integrated into the middleware) and customizable load balancing strategies. Implementation of a low level service requires in order to stay efficient to avoid the necessity for high overhead. DLBS aims to be a generic load balancing service and an easy and efficient portability with a CORBA object request broker has been possible thanks to aspect oriented programming.