快取是高速設備和低速設備間的橋樑。本身的目的在於在有較好反應時間的設備上儲存資料，避免向較低速設備存取，如L1/L2 cache、檔案快取、proxy cache等。隨著網路的蓬勃發展，如Gigabit, Myrinet, InfiniBand, NPU等，網路傳輸速度日益飛快，而硬碟等機械裝備的速度雖然也有進步，但和高速網路的進步想比實在有限，而硬碟存取常常造成整體系統效率的瓶頸。將資料透過高速網路快取在其他電腦記憶體內，比從硬碟直接讀取來的有效率。因此，這篇論文主在闡述如何在一個由PC所組成的叢集式系統中，設計並實作一廣域的快取管理系統。我們的目的在使用位於作業系統中的一個單一、一致且分散式的記憶體快取管理中介軟體。藉由此中介軟體，應用程式能從叢集式系統的記憶體得到優勢且能獲得更多可使用的記憶體，即使是由不同的作業系統所組成的叢集式系統。此中介軟體整合了整個叢集式系統的快取記憶體（在此指檔案快取)，對於應用程式而言，看到的只是一個大型而唯讀的快取空間。所有的應用程式透過此系統分享檔案並進行快取，不需要再自己維護快取，而避免重複複製。我們已經在一個由Gigabit高速網路所連接的PC叢集式系統上實作出此一中介軟體，而其作業系統包含Linux和Microsoft Windows NT。在實驗中，我們實際測量一個擁有大量檔案的網頁伺服器效能，結果因避免掉了大部分的磁碟機讀取，成功的提升了網頁伺服器的效能，隨著叢集式系統節點的增加，效率可呈1到2.5倍的成長。

Cache is the bridge among high-speed equipment and low-speed equipment to fetch quickly. One's own purpose lies in storing the materials in having better equipment of time of response, prevent from to low-speed equipment deposit and withdraw relatively, for instance L1/L2 cache , file fetching , proxy cache ,etc. soon. With the flourishing development of the network, such as Gigabit, Myrinet, InfiniBand, NPU ,etc., the transmission speed of the network is day by day quickly, though the speed that and the machinery , such as hard disk ,etc. are equipped makes progress, but and the progress of the high-speed network wants to be more limited than really, and the hard disk deposits and withdraw the bottleneck often causing whole systematic efficiency. Fetch the materials in the storing device of other computers quickly through the high-speed network, more efficient than the ones that read and fetched more directly from the hard disk. So, the thesis in explain main fact how in one made up by PC system of gathering together etc., design and make fetching the administrative system quickly of a wide area in fact. Our purpose fetches and manages intermediary's software quickly in the storing device of a single , unanimous and dispersing type of using lying in the operating system. With this intermediary software, the application program can receives the advantage from the gathering together type systematic storing device and can get more storing device that can be useds , even a gathering together type system made up by different operating systems. This intermediary software combine whole of person who gather together system fetch person who store soon (mean file fetch soon here), as to application program, it was only a large-scale and only read fetching the space quickly that was seen. All application program share the file and fetch quickly through this system, do not need to safeguard fetching quickly by oneself again , but avoid duplicating repeatedly. We have already made this intermediary software in fact in a PC gathering together type system joined by Gigabit high-speed network, and its operating system includes Linux and Microsoft Windows NT. In the experiment, a webpage server efficiency of having a large number of files of our actual measurement, having avoided losing most disk drives to read in because the result, the efficiency of the successful improvement webpage server, with the gathering together type increase with nodal system, efficiency can present growth of 1 to 2.5 times .

Bibliography

[1] T. Anderson, D.Culler, and D. Patterson. A Case for NOW (Networks of Workstations). In IEEE Micro, 15(1):54-64, 1995

[2] Thomas Anderson, Michael Dahlin, Jeanna Neefe, David Patterson, Drew Roselli, and Randolph Wang. Serverless network file systems. In Proceedings of the 15th Symposium on Operating system Principles. ACM, pages 109-126, Copper Mountain Resort, Colorado, December 1995

[3] Marshall Bern, Daniel Greene, and Arvind Raghunathan. Online algorithms for cache sharing. In Proceedings of the twenty-fifth annual ACM symposium on Theory of computing, pages 422-430. ACM Press, 1993.

[4] B. Bloom. Space/time trade-offs in hash coding with allowable errors, Communications of the ACM, vol 13, no, 7, pages 422-426, July 1970.
[5] Anawat Chankhunthod, Peter B. Danzig, Chunk Neerdaels, Michael F. Schwartz, and Kurt J. Worrell. A hierarchical internet object cache. In USENIX Annual Technical Conference, pages 153-164, 1996.

[6] G. Chen, C.L. Wang, and F.C.M. Lau. Building a scalable web server with global object space support on heterogeneous cluster. In Proceeding of IEEE International Conference on Cluster Computing, 2001.

[7] T. Cortes, S. Girona, and J. Labata. PACA: A Distributed File System Cache for Parallel Machines. Technical Report UPC-DAC-1995-20, Department d’Arquitectura de computadors, Universitat Politecnica de Catalunya (UPC), June 15 1995.

[8] Michael Dahlin, Randolph Wang, Thomas E. Anderson, and David A. Patterson. Cooperative caching: Using remote client memory to improve file system performance. In Operating System Design and Implementataion, pages 267-280, 1994.

[9] Li Fan, Pei Cao, Jussara Almeida, and Andrei Z. Broder. Summary cache: a scalable wide area Web cache sharing protocol. IEEE/ACM Transactions on Networking, 8(3):281-293, 2000.

[10] Michael J. Feeley, William E. Morgan, Frederic H. Pighin, Anna R. Karlin, Henry M. Levy, and Chandramohan A. Thekkath. Implementing global memory management in a workstation cluster. In Symposium on Operating Systems Principles, pages 201-212, 1995.

[11] Thu D. Nguyen Francisco Matias Cuenca-Acuna. Cooperative caching middleware for cluster-based servers. In Tenth IEEE International Symposium on High Performance Distributed Computing (HPDC-10). IEEE Press, August 2001.

[12] Syam Gadde, Jeff Chase, and Michael Rabinovich. A taste of Crispy Squid. In Proceedings of the Workshop on Internet Server Performance (WISP’98), 1998.

[13] Zornitza Genova and Kenneth J. Christensen. Challenges In URL switching for implementing globally distributed web sites. In ICPP Workshop, pages 89-94, 2000.

[14] John H. Hartman and John K. Ousterhout. The Zebra striped network file system. In Proceedings of the 14th ACM Symposium on Operating Systems Principles, pages 29-43, 1993.

[15] Josh H. Howard, Michael L. Kazar, Sherri G. Menees, David A. Nichols, M. Satyanarayanan, Robert N. Sidebotham, and Michael J. West. Scale and performance in a distributed file system. ACM Transactions on Computer Systems, 6:51-81, 1988.

[16] ICP working group. National Lab for Apllied Network Research.
http://www.irache.net/.
[17] InfiniBand Trade Association
http://www.infinibandta.org/.
[18] Information and Communication Technologies at CSIRO Australia.
http://www.ict.csiro.au/gigabit/gigabitradio.htm.

[19] David Karger, Eric Lehman, Tom Leighton, Mathhew Levine, Daniel Lewin, and Rina Panigraphy. Consistent hashing and random trees: Distributed caching protocols for retrieving hot spots on the world wide web. In ACM Symposium on Theory of Computing, pages 654-663, May 1997

[20] Evangelos P. Markatos and George Dramitinos. Implementation of a reliable remote memory pager. In USENIX Annual Technical Conference, pages 177-190, 1996.

[21] Myrinet. http://www.myrinet.com/.

[22] NTT Microsystem Integration Laboratories. Development of Large-Scale Integrated Optical Switched Based on Silica Waveguide.
http://www.ntt.co.jp/saiyo/e/rd/review/pdf/ct07.pdf.

[23] S. Rhea and W. Weimer. Data location in the oceanstore, 1999.

[24] Sean C. Rhea and John Kubiatowicz. Probabilistic location and routing. In Proceedings of INFOCOM 2002, 2002.

[25] Prasenjit Sarkar and John H. Hartman. Hint-based cooperative caching. 18(4):387-419, 2000.

[26] Bill N. Schilit and Dan Duchamp. Adaptive remote paging for mobile computers. Technical Report CUCS-004-91, Department of Computer Science, Columbia University, 1991.

[27] Standard Performance Evaluation Corporation.
The specweb99 benchmark.
http://www.spec.org/osg/web99/.

[28] Chandramohan A. Thekkath, Timothy Mann, and Edward K. Lee. Frangipani: A scalable distributed file system. In Symposium on Operating Systems Principles, pages 224-237, 1997.

[29] Jeffery Westbrook. Randomized algorithms for multiprocessor page migration. In DIMACS Series in Discrete Mathematics and Theoretical Computer Science, volume 7, pages 135-150, 1992.

[30] Youhui Zhang, Weimin Zheng. User-Level Commnuication based Cooperative Caching. In ACM SIGOPS Operating Systems Review, volume 37, pages 23-33, January 2003