The Flash crowds are rapid increase in access to contents of web sites, which makes the web sites inaccessible, leaving the clients with unsatisfied requests. The major shortcoming of flash crowds researches is that they do not assist vital resizing feature of a cloud of the surrogates; the surrogates involved in the alleviation process do not change from the start to the end of flash crowds. Our system, FCAN (Flash Crowds Alleviation Network) is a system to provide resources to web sites to overcome flash crowds. A main feature of FCAN is its dynamically resizing feature, which can adapt to request load of flash crowds by enlarging or shrinking a cloud of surrogate servers used by the web sites. In this paper, we present a new feature of FCAN to support multiple servers which experience different flash crowds simultaneously, and show experiment results with real web log data provided by Live Eclipse 2006.

Abstract. Recently, attention has been focused on incomplete-type spinal cord injury to the central spine caused by pressure on parts of the white matter conducting pathway such as the pyramidal tract. In this paper, we focus on a training robot designed to assist with primary walking pattern training. The target patient of this training robot is one who is relearning the basic functions of the usual walking pattern, especially for patients with the incomplete type SCI to the central spine, who are capable of standing by themselves but are not capable of performing walking motions. From the perspective of human engineering, we monitored the operator's actions to the robot and investigated the movement of joints of the lower extremities, the circumference of the lower extremities and the exercise intensity with the machine. The concept of the device is to provide mild training without any sudden changes in heart rate or blood pressure, which will be particularly useful for the elderly and disabled.

SUMMARYMost Internet service providers offer services using multiple servers in order to avoid service congestion. In such environments, problems have arisen in regard to the client not knowing which server to choose. This has led to the need for a server selection mechanism that will automatically select the optimal server. In our paper, we first identify the requirements of a server selection mechanism, and then propose a novel mechanism that provides a satisfactory response to them. The mechanism has an extensive range of usability as a result of being based on a Domain Name System (DNS), and enables a diversity of server selection policies to be integrated with it. These server selection policies also allow decisions to be made based on roundtrip times and other network status parameters. Allowing selection policies to be changed dynamically ensures compatibility with a variety of conditions. We used a prototype system to conduct assessment experiments on a group of widely distributed servers and to show the effectiveness of our server selection mechanism.

No abstract

No abstract