[Show abstract][Hide abstract] ABSTRACT: . The buffer cache management algorithm used by almost all operating systems is known as LRU(Least Recently Used). LRU is known to perform well in applications with irregular reference pattern. However, it shows poor performance for sequential-pattern or looping-pattern applications with large-size data. For sequential or looping reference pattern, MRU(Most Recently Used) replacement policy is frequently used to improve the buffer cache performance. On the way of implementing MRU on Linux, we observed that MRU shows lower response time by up to 100% compared to LRU. Indirect blocks, which are used in the file structure of Unix family operating systems for large-size file, are the main reason of decreasing performance. Indirect blocks are fetched but immediately replaced by MRU replacement policy, even those will be soon and frequently needed again. Based on this observation, we propose a buffer replacement policy named `LMRU'. LMRU maintains frequently-used blocks such as indirect bl...
[Show abstract][Hide abstract] ABSTRACT: Caching and prefetching are well-known solutions to the problem of long latency in both wired and wireless networks. Caching is prevalently used to reduce round-trips and network traffic. The prefetching and uploading-prefetching data from fixed hosts on a wired network and uploading to mobile hosts-in the base station for the WWW has been proposed to solve those problems and has shown high performance improvement. The performance of prefetching in the base station can be improved when the base station has the caching information of mobile hosts. Cache-index forwarding can be used to let the base station see information of caching on mobile hosts. Here we propose a method named “CINDEX: Cache-Index Forwarding per Document for the WWW”, which sends the information of caching data on mobile hosts per document to the base station to eliminate unessential data prefetching and uploading
Parallel and Distributed Systems, 1998. Proceedings., 1998 International Conference on; 01/1998
[Show abstract][Hide abstract] ABSTRACT: Linux Virtual Server (LVS) is highly scalable and available server built on a cluster of real servers. For offering the successful service, there are many scheduling methods in LVS. But traditional scheduling methods are not offered sophisticated load balancing because it is regardless of the real load information of real servers. Agent-based scheduling method is considered the real load information of real servers such as the presenting in 'Bit Project No.61'. But load balancer has a lot of additional communication overhead . We propose ABSS: Agent-based Sophisticated and Scalable Scheduling Method which can be considered the real load information for sophisticated load balancing and be reduced the additional communication overhead of load balancer by using the itinerant load information packet.