為了節省重複傳送熱門資料所佔用的頻寬及加速資料之取得，快取伺服器 (cache server) 已被廣泛地應用。快取伺服器的位置會影響命中率(hit rate)及取得資料之回應時間，且各快取伺服器間內容之重複性過高，及使用者可能設定錯誤以致效能不彰等均是急待解決的問題。
近來主動式網路(active networks)的觀念被提出來，其與傳統網路之不同處是每個節點均可根據封包內容執行簡單的程式碼以處理經過的封包，藉以分擔客戶端的負載，並減少網路流量。一般可將主動式節點之主記憶體劃出一部份作為快取空間，而快取空間再劃分為可用來暫存程式碼的程式碼快取區域，及暫存所流經資料的資料快取區域，因此每個主動式節點皆可視為擁有一快取伺服器，但由於此類型快取伺服器之容量有限，若資料被快取的頻率過高，將造成空間的浪費；反之則增加延遲時間。因此如何組織這些由主動式節點所組成的分散式快取伺服器，使之能各自對資料進行快取與否之決策，已成一重要的研究課題。
本論文提出一可依據使用者容忍的回應時間及網路環境之變化來進行快取資料與否之決策演算法。以使用者容忍的回應時間為依據，利用快取伺服器的自我管理(Self management)以決定在不同的網路狀況下資料被快取的頻率；再利用快取伺服器的自我組織(Self organization)將資料均勻地分散儲存於沿途的快取伺服器中，以提高使用者在容忍的回應時間內取到資料的機率，也由於資料是透通地(transparently)在所經過的路徑上被處理，如此可避免使用者的設定錯誤以致效能不彰。由模擬的結果顯示，相較於固定距離快取一次的方法，本方法重複的檔案數目較高約12%，但是命中率提高了1.2%，在使用者容忍時間內拿到資料的機率，本方法約提高了7%。而在資料傳輸的延遲時間方面，平均約降低了19%。

The cache server has been widely applied to save the bandwidth for popular data, and to speed up the acquirement for data. The location of the cache servers influences the hit rate and the response time. The problems for the high duplicated contents among cache servers, and the degradation on the efficiency caused by improper configurations from users have become important issues.
Recently the concept of active network has been proposed, difference where each node in the active network can execute simple code according to the content of packets to process the passed-through packets. Thus the load of client sides cache shared, and the traffic of network is reduced. Generally the cache area the cache server is located in the main memory of an active node. The cache area is divided into the code cache storing code temporarily and the data cache storing the passed-through data. Therefore, each active node can be treated as a cache server. For the capacity the main memory is finite, the cache area has to be arranged appropriately. If the frequently, the cache area may be utilized inefficiently; otherwise, the response time may be increased. Therefore, how to appropriately organize these distributed cache servers in active networks and to decide whether to cache data or not are addressed in the thesis.
The thesis proposes a caching algorithm to decide whether cache data or not, according to the acceptable response time of users and the change of the network environment. Based on the acceptable response time, The proposed algorithm is capable of determining the frequency of the cached data for various network environment by self-mangement of cache servers. The data is stored fairly among cache servers by self-organizing the cache servers. Therefore, the probability that users get the data before the acceptable response time is increased. For the data is processed transparently along the passed-through path, the degradation of the efficiency due to users’ improper configuration can be obviously prevented. Compared with the method that caches data once within a fixed distance, the proposed algorithm is able to increase the number of data copies by 12%,but increase the hit rate by 1.2%, increase the probability that users get the data before the acceptable response time by 7%, and the response time is decreased by 19%.

[1]Squid internet object cache. http://squid.nlanr.net/Squid/.
[2]Cisco Cache 500 series. http://www.cisco.com/warp/public/cc/cisco/mkt/scale/
cache/index.shtml
[3]D. Wessels and K. Claffy, “Internet Cache Protocol (ICP), version 2,” Network Working Group RFC 2186, Sept. 1997.
[4]A. Rousskov and D. Wessels, “Cache Digests,” Computer Networks, vol.30 pp.2155-2168, Nov. 1998
[5]V. Valloppillil and K. W. Ross, “Cache Array Routing Protocol v1.0,” Internet Draft, http://ircache.nlanr.net/Cache/ICP/draft-vinod-carp-v1-03.txt, Feb. 1998.
[6]S. Bhattacharjee, K. Calvert and E. Zegura, “Self-Organizing Wide-Area Network Caches,” Proceedings of IEEE INFOCOM ''98, San Francisco, CA, USA, vol.2 pp.600-608, Mar. 1998.
[7]D. Tennenhouse, J. Smith, W. Sincoskie, D. Wetherall and G. Minden, “A Survey of Active Network Research,” IEEE Communications Magazine, vol. 135, no. 1, pp. 80-86, Jan. 1997.
[8]Wetherall and D. Tennenhouse. “The Active IP Option,” Proceedings of the 7th ACM SIGOPS European Workshop, Connemara, Ireland, Sept. 1996.
[9]E. Nygren, “The Design and Implementation of a High Performance Active Network Node,” MIT Master Thesis, Massachusetts, Feb. 1998.
[10]D. Alexander, W. Arbaugh, M. Hicks, P. Kakkar, A. Keromytisk, J. Moore, C. Gunter, S. Nettles and J. Smith, “The SwitchWare Active Network Architecture,” IEEE Network Magazine, vol. 12, no. 3, pp. 29-36, May/June 1998.
[11]Caml home page. http://pauillac.inria.fr/caml/index-eng.html
[12]D. Alexander W. Arbaugh, A. Keromytis and J. Smith, “A Secure Active Network Environment: Realization in SwitchWare,“ IEEE Network Magazine, vol. 12, no. 3, pp. 37-45, May/June 1998.
[13]B. Schwartz, A. Jackson, W. Strayer, W. Zhou, R. Rockwell and C. Partridge, “Smart Packets for Active Networks,” http://www.net-tech.bbn.com/smtpkts/
smart.ps.gz., Jan. 1998.
[14]S. Bhattacharjee, K. L. Calvert and E. W. Zegura, “On Active Networking and Congestion,” Technical Report GIT-CC-96/02, College of Computing Georgia Institute of Technology, Georgia, 1996.
[15]T. Faber, “ACC: Using Active Networking to Enhance Feedback Congestion Control Mechanisms,” IEEE Network, vol. 12, no. 3, pp. 61-65, May/June 1998.
[16]M. Calderon, M. Sedano, A. Azcorra and C. Alonso, “Active Networks Support for Multicast Applications,” IEEE Network, vol. 12, no. 3, pp. 46-52, May/June 1998.
[17]L. H. Lehman, S. J. Garland and D. L. Tennenhouse, “Active Reliable Multicast,” IEEE INFOCOM ’98, San Francisco, USA, vol.2 , pp. 581-589, Apr. 1998.
[18]R. Wittmann and M. Zitterbart, “AMnet: Active Multicasting Network,” IEEE International Conference on Communication, vol. 2, pp. 896-900, 1998.
[19]Amit Kulkarni and Gary Minden, “Composing Protocol Frameworks for Active Wireless Network,” IEEE Communications Magazine, vol 3, pp.130-137, March 2000.
[20]ABONE status home page. http://sequoia.csl.sri.com:7000/java/abonestat.html
[21]D. L. Tennenhouse, S. J. Garland, L. Shrira and M. F. Kaashoek, “From Internet to ActiveNet,” http://www.sce.carleton.ca/netmanage/activeNetworks/rfc96.html, Jan. 1996.
[22]D. Scott Alexander, B. Braden, C. Gunter, A. Jackson, A. Keromytis, G. Minden and D. Wetherall, “Active Network Encapsulation Protocol (ANEP),” Experimental RFC draft, July 1997
[23]B. Braden, M. Hicks and C. Tschudin, “Active Network Overlay Network (ANON),” Experimental RFC draft, Dec. 1997
[24]Steven Glassman. A caching relay for the world wide web. In First International Conference on the World Wide Web, CERN, Geneva, Switzerland, May 1994. http://www1.cern.ch/WWW94/PrelimProcs.html.
[25]E. Zegura and K. Calvert. “Georgia Tech Internet Topology Models,” http://www.cc.gatech.edu/projects/gtitm.