DTS: Broadcast-based Content-aware TCP Connection Handover
作者： Hajime Fujita , Yutaka Ishikawa
关键词: Code (cryptography) 、 Node (networking) 、 Computer science 、 Computer network 、 Linux kernel 、 Operating system 、 Benchmark (computing) 、 Consistent hashing 、 Throughput 、 Handover 、 Cluster Node
摘要: In this paper we propose DTS (Distributed TCP Splicing), a new mechanism for performing content-aware connection switching in broadcast-based single IP address cluster. Broadcast-based design enables each cluster node to continue provide services clients even when other nodes the fail. Each request from client is first distributed among using consistent hashing method, order share inspection workload. Then transferred an appropriate according content of request. implemented on Linux kernel module and does not require any modification main code, server applications, or applications. With 8-node configuration, with multiple inspectors achieves about 3.4 times higher throughput compared inspector configuration. A SPECweb 2005 Support benchmark also conducted four cluster, where reduces total amount disk accesses locality-aware distribution almost halves number file downloads that fail meet speed requirement.
ceek.jp参考文章(13)
Mike Keefe, Patrick O'Rourke, Performance Evaluation of Linux Virtual Server usenix large installation systems administration conference. pp. 79- 92 ,(2001)
Alan Robertson, Linux-HA heartbeat system design ALS'00 Proceedings of the 4th annual Linux Showcase & Conference - Volume 4. pp. 20- 20 ,(2000)
D. Kerdlapanan, A. Khunkitti, Content-based load balancing with multicast and TCP-handoff international symposium on circuits and systems. ,vol. 2, pp. 348- 351 ,(2003) , 10.1109/ISCAS.2003.1205979
Valeria Cardellini, Emiliano Casalicchio, Michele Colajanni, Philip S. Yu, The state of the art in locally distributed Web-server systems ACM Computing Surveys. ,vol. 34, pp. 263- 311 ,(2002) , 10.1145/508352.508355
David Karger, Eric Lehman, Tom Leighton, Rina Panigrahy, Matthew Levine, Daniel Lewin, Consistent hashing and random trees: distributed caching protocols for relieving hot spots on the World Wide Web symposium on the theory of computing. pp. 654- 663 ,(1997) , 10.1145/258533.258660
Hiroya Matsuba, Yutaka Ishikawa, Single IP Address Cluster for Internet Servers international parallel and distributed processing symposium. pp. 1- 10 ,(2007) , 10.1109/IPDPS.2007.370255
Vivek S. Pai, Mohit Aron, Gaurov Banga, Michael Svendsen, Peter Druschel, Willy Zwaenepoel, Erich Nahum, Locality-aware request distribution in cluster-based network servers architectural support for programming languages and operating systems. ,vol. 33, pp. 205- 216 ,(1998) , 10.1145/291006.291048
Manish Marwah, Shivakant Mishra, Christof Fetzer, Fault-tolerant and scalable TCP splice and web server architecture 2006 25th IEEE Symposium on Reliable Distributed Systems (SRDS'06). pp. 301- 310 ,(2006) , 10.1109/SRDS.2006.21
O. Spatscheck, J.S. Hansen, J.H. Hartman, L.L. Peterson, Optimizing TCP forwarder performance IEEE ACM Transactions on Networking. ,vol. 8, pp. 146- 157 ,(2000) , 10.1109/90.842138
Peter Druschel, Mohit Aron, Darren Sanders, Willy Zwaenepoel, Scalable content-aware request distribution in cluster-based networks servers usenix annual technical conference. pp. 26- 26 ,(2000)