C. Staelin and H. Garcia-Molina. File system design using large memories. In Proceedings of the Fifth Jerusalem Conference on Information Technology (JCIT), pages 11-- 21. IEEE, Oct. 1990.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....to ameliorate this cost [Blaze and Alonso 1992, Griffioen and Appleton 1994, Grimsrud et al. . 1993, Howard et al. . 1988, Korner 1990, Kotz and Ellis 1990, Kroeger and Long 1996, Lazowska et al. . 1986, Lei and Duchamp 1997, Muntz and Honeyman 1992, Palmer and Zdonik 1991, Schroeder et al. . 1985, Staelin and Garcia Molina 1990, Tait and Duchamp 1991, Vitter and Krishnan 1996] These schemes generally interpose an agent between the application software and the file system. This agent observes and records the disk access patterns of the applications. At a later time, when the beginning of a pattern is repeated, the agent ....

....to hoarding (Section 5.3.4, p. 55) Ideally, the weight would based on the cost of moving the entire cluster, which would be affected both by the sizes of other members and by which members were already present in the local hoard. Cluster Temperatures. Some other researchers [Salem et al. . 1992, Staelin and Garcia Molina 1990] have suggested using a temperature to characterize the activity or popularity of a cache object. The idea is that the temperature rises higher (usually in a linear fashion) whenever an object is accessed, and decays over time, usually exponentially. An object that is used frequently will gain ....

Carl Staelin and Hector Garcia-Molina. "File System Design using Large Memories." In Proceedings of the Fifth Jerusalem Conference on Information Technology, pp. 11--21. IEEE, 1990.

....a step would help more than hurt. Our algorithm is implicitly based on whole file caching. This is perhaps acceptable, since studies indicate that the great majority of files are read in their entirety [14, 39] Furthermore, there is a minor trend toward file systems that use whole file caching [21, 54]. Very large files and shared files should remain at the server, however, and we have not yet addressed the problem of how to deal with these files. Finally, we would like to gather more trace data from different environments. The traces that we are currently using, however, provide encouraging ....

C. Staelin and H. Garcia-Molina. File System Design Using Large Memories. Technical Report CS-TR-246-90, Princeton University, June 1990.

....of its files are accessed. Our simulator is implicitly based on whole file caching. This is perhaps acceptable, since studies indicate that the great majority of files are read in their entirety [4, 7] Furthermore, there is a minor trend toward file systems that use whole file caching 13 [5, 10]. Very large files and shared files should remain at the server, however, and we have not yet addressed the problem of how to deal with these files. Finally, we would like to gather more trace data from different environments. The traces that we are currently using, however, provide encouraging ....

C. Staelin and H. Garcia-Molina. File System Design Using Large Memories. Technical Report CS-TR-246-90, Princeton Univ., June, 1990.

....of various clustering algorithms. Adding the clustering algorithm to iPcress improved disk performance 10 30 . Some of the material presented in this dissertation has been published previously. Material from Chapter 3 appeared in Reference [110] Chapters 4 and 5 contain material from References [112] and [113] and Chapter 7 contains material from Reference [111] Chapter 2 Prior Work In order to design a high performance file system, we must first understand how existing file systems are used, so we analyze file access patterns in existing systems. We describe several file systems and ....

C. Staelin and H. Garcia-Molina. File system design using large memories. In Proceedings IEEE Jerusalem Conference, pages 11--21, Jersusalem, Oct. 1990.

....are proposing to automate this process, and we evaluate expected performance gains. We also stress that this reorganization will be done frequently, possibly even on a continuous basis. The development of iPcress consists of two concurrent efforts. One is the actual implementation of the system [2]. A current prototype is running, implementing a large file cache with variable sized blocks both in memory and on secondary storage. It allows multiple devices within a single file system. The second effort is a performance evaluation of the various planned optimizations, in order to guide the ....

C. Staelin and H. Garcia-Molina, "File system design using large memories, " in Proceedings Fifth Jerusalem Conference on Information Technology, Oct. 1990.

No context found.

C. Staelin and H. Garcia-Molina. File system design using large memories. In Proceedings of the Fifth Jerusalem Conference on Information Technology (JCIT), pages 11-- 21. IEEE, Oct. 1990.

No context found.

Carl Staelin and Hector Garcia-Molina. File System Design Using Large Memories. Technical Report CS-TR-246-90, Princeton University, Department of Computer Science, February1990. Appeared in 5th JerusalemConferenceon Information Technology, Jerusalem Israel, Oct 1990.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC