Introduction to Content Switch.ppt

《Introduction to Content Switch.ppt》由会员分享，可在线阅读，更多相关《Introduction to Content Switch.ppt（32页珍藏版）》请在麦多课文档分享上搜索。

1、Introduction to Content Switch,C. Edward Chow Department of Computer Science University of Colorado at Colorado Springs,12/22/2000,Edward Chow ACSD Project Status 2,Outline of the Talk,What is a Content Switch? What Services it Can Provide Content Switch Example Related Technologies Content Switch A

2、rchitecture and Basic Operations TCP Delay Binding and Related Improvement Content Switch Rule and Conflict Detection Related Load Balancing Research Results,12/22/2000,Edward Chow ACSD Project Status 3,Content Switch (CS),Route packets based on high layer (Layer 5/7) headers and content. Examples:

3、Direct Web traffic based on pattern of URLs, host tags, cookies. Can Route incoming email based on email address; Connect POP/IMAP based on login Web switches and Intel XML Director/accelerator are special cases of content switch.,12/22/2000,Edward Chow ACSD Project Status 4,What Services It Can Pro

4、vide,Enabling premium services for e-commerce, ISP, and Web hosting providers Load Balancing and High Available Server Clusters: Web, E-commerce, Email, Computing, File, SAN Policy-based networking, differential/QoS services. Firewall, Strengthening DoS protection, cache/firewall load-balancing Flas

5、h-crowd management Email Spam Protection, Virus Detection/Removal Applet Authentication/Filtering,12/22/2000,Edward Chow ACSD Project Status 5,F5 VRM Solution,12/22/2000,Edward Chow ACSD Project Status 6,Intel Netstructure XML Director 7280,Example of Rule: Server1: create */order.asp & /AmountValue

6、 = 10000,12/22/2000,Edward Chow ACSD Project Status 7,Phobos In-Switch,Only load balancing switch in a PCI card form factor Plugs directly into any server PCI slot Supports up to 8,192 servers, ensuring availability and maximum performance Six different algorithms are available for optimum performan

7、ce: Round Robin, Weighted Percentage, Least Connections, Fastest Response Time, Adaptive and Fixed. Provides failover to other servers for high-availability of the web site U.S. Retail $1995.00,12/22/2000,Edward Chow ACSD Project Status 8,E-Commerce Example: 1. Client,Client submits via HTTP/Post (o

8、r SOAP) the following purchase in XML:CCL111222333309121544IBM Thinkpad T2150001050000309121538Intel wireless LAN PC Card20010200052000 ,12/22/2000,Edward Chow ACSD Project Status 9,E-Commerce Example: 2. Content Switch,Content switch receives the packet. Recognize it is a http post request from htt

9、p request line POST /purchase.cgi HTTP/1.1 Recognize it is an XML document from the meta header content-type: TEXT/XML Parsing XML content Extract values of tag sequences: 52000 purchase/totalAmount CCL purchase/customerName Rule 1 is matched and packet is routed to one of highSpeedServers. Rule 1:

10、if (xml.purchase/totalAmount 5000) routeTo(highSpeedServers); Rule 2: if (xml.purchase/customerName = CCL) routeTo(specialCustomerServers);,12/22/2000,Edward Chow ACSD Project Status 10,No Free Lunch: Penalty of Having Content Switch, Increased packet processing time. For XML Director/Accelerator, i

11、t needs to parse XML document and match tag sequences. 1-3? order of processing time,12/22/2000,Edward Chow ACSD Project Status 11,Related Technologies,Application level solution: Proxy server; Apache/Tomcat/Servlet; Microsoft NLB Kernel level layer 4 load balancing solution: http:/www.linuxvirtuals

12、erver.org/ Joseph Marks presentation LVS-NAT(Network Address Translation) web page LVS-IP Tunnel web page LVS-DR (Direct Routing) web page Hardware solution: Cisco 11000, F5 (Big IP), Alteon Web Systems, Foundry Networks (ServerIron), Good information: Foundry ServerIron Installation and Configurati

13、on Guide, May 2000.,12/22/2000,Edward Chow ACSD Project Status 12,Basic Operations of Content Switching,CS Rule Matching Algorithm,Header Content Extraction,Packet Classification,CS Rules,Packet Routing (Load Balancing),CS Rule Editor,Incoming Packets,Forward Packet To Servers,Network Path Info,Serv

14、er Load Status,CS: Content Switching,12/22/2000,Edward Chow ACSD Project Status 13,Content Switch Architecture,Apostolopoulos 2000,12/22/2000,Edward Chow ACSD Project Status 14,Efficient Software Architecture,Tasks: Million Packets with thousand of rules to match and load balancing algorithms to run

15、. How to assign tasks to the processors and threads? Packet Extraction (Understand header formats, XML parsing) Content Switching Rule Matching Packet Routing (Load Balancing, Bandwidth Control) How Much Packet Processing Should Controllers Do? What a controller can do? A Typical Parallel Processing

16、 Problem?,12/22/2000,Edward Chow ACSD Project Status 15,TCP Delay Binding,client,content switch,server,step1,step2,SYN(CSEQ),SYN(DSEQ),ACK(CSEQ+1),step4,step9,step10,step5,step6,SYN(CSEQ),SYN(SSEQ),ACK(CSEQ+1),step8,DATA(CSEQ+1),ACK(SSEQ+1),DATA(SSEQ+1),ACK(CSEQ+lenR+1),DATA(DSEQ+1),ACK(CSEQ+LenR+1)

17、,ACK(DSEQ+,lenD+1),ACK(SSEQ+lenD+1),lenR: size of http request.,lenD: size of return document,.,step7,ACK(SSEQ+1),step11,12/22/2000,Edward Chow ACSD Project Status 16,Lesson Learned in Implementing TCP Delay Binding,In our Linux 2.2 kernel-based content switch prototype, we found client sends duplic

18、ate requests after step 3. It overloads the content switch and the real server. Reason: Client TCP time-out, retransmit Content switch printk() overhead, too many debug msgs It could happens when there are many content rules, slow server response. Solution: content switch sends ack(CSEQ+LenR+1) to s

19、top retransmit.,12/22/2000,Edward Chow ACSD Project Status 17,Pre-Allocate Server Scheme,client,content switch,Pre-allocated server,.,Guess routing decision based on IP/Port#/HistoryAdvantage: Faster than TCP delay binding.Possible direct route between client and serverReduce session processing over

20、head no need to convert server sequence #,12/22/2000,Edward Chow ACSD Project Status 18,Degenerated to TCP Delay Binding If Guess Wrong,client,content switch,Pre-allocated server,step1,step2,SYN(CSEQ),SYN(SSEQ),ACK(CSEQ+1),DATA(CSEQ+1),ACK(SSEQ+1),step3a,SYN(CSEQ),SYN(SSEQ),ACK(CSEQ+1),step4,step5,D

21、ATA(RSEQ+1),ACK(CSEQ+lenR+1),DATA(SSEQ+1),ACK(CSEQ+LenR+1),ACK(DSEQ+,lenD+1),ACK(SSEQ+lenD+1),FIN(CSEQ+1),Right server,Sequence # conversion needed,12/22/2000,Edward Chow ACSD Project Status 19,Filter Process Scheme,Filter Process run on server,client,content switch,server,step1,step2,SYN(CSEQ),SYN(

22、DSEQ),ACK(CSEQ+1),DATA(CSEQ+1),ACK(DSEQ+1),step3,step4,a,step5,step6,step7,step8,SYN(CSEQ),SYN(SSEQ),ACK(CSEQ+1),DATA(CSEQ+1),ACK(SSEQ+1),DATA(SSEQ+1),ACK(CSEQ+lenR+1),DATA(DSEQ+1),ACK(CSEQ+LenR+1),ACK(DSEQ+,lenD+1),ACK(SSEQ+lenD+1),step4b,Migrate (Data, CSEQ, DSEQ),12/22/2000,Edward Chow ACSD Proje

23、ct Status 20,Multiple HTTP Requests from One TCP Connection,A keep alive TCP connection may include multiple HTTP “GET” requests.Content Switch examines each “GET” request and makes new routing decision.Content Switch establishes another connection with a different server based on the routing decisi

24、on.Those HTTP responses from different servers need to be interleaved and seen by the user as if from the same server.Solutions: In order delivery (buffer requirement); Out of order delivery (seq# tracking)?Problems: Should we throw away earlier html requests if receive later requests?,. .,.,client,

25、NAT approach,uccs.jpg,rocky.mid,home.htm,Index.htm,Content Switch,server1,server2,server9,12/22/2000,Edward Chow ACSD Project Status 21,Multiple HTTP Requests from One TCP Connection,Can servers return documents directly to client in keep-alive session case? Can equivalent VS-Tunnel or VS-DR be impl

26、emented using Content Switch?,. .,.,client,uccs.jpg,rocky.mid,home.htm,Content Switch,server1,server2,server9,12/22/2000,Edward Chow ACSD Project Status 22,Content Switch Rule Survey,Survey shows that existing switches support rules in basic (condition action) or (action condition) form some define

27、condition as class, then specify the action in separate statement or command simple single conditional term command line interface (to facilitate incremental update?) Actions can include reject, forward, put in queue (for bandwidth control, scheduling),12/22/2000,Edward Chow ACSD Project Status 23,C

28、ontent Switch Rule Design,Rule syntax generic to support all Intended features. Use simple C if statement syntax rule: if (condition) action Easy to read Allow optimization using c compiler Condition consists of multiple terms of variable relational_operator value e.g. xml.purchase/totalAmount 50000

29、 smtp.to = “chowcs.uccs.edu” cookie.name = “servlet1” bitmatch(64, 8, 0xff) = 64 # above mean TTL=64 idea from netfilter universal filter suffix(variable, string) e.g. suffix(url, “gif”) regex(variable, pattern) e.g. regex(url, “/purchase”) Action consists of reject, forward(server| queue) loadBalan

30、ce(serverGroup, loadBalancingAlgorihtm),12/22/2000,Edward Chow ACSD Project Status 24,Efficient CS Rule Matching,Brute force, strict priority: Rules are executed in sequential manner. Efficient Rule Matching Method: Organize Rules so that rules can be skipped based on existing content types. Utilize

31、 compiler optimization technique.,12/22/2000,Edward Chow ACSD Project Status 25,Simple CS Rule Editor GUI,12/22/2000,Edward Chow ACSD Project Status 26,Conflict Detection on Content Switching Rules,Detect conflicts among rules or rule set. Absolute conflict type: r1: if (xml.purchase/customerName =

32、“CCL”) routeTo(r1) r2: if (xml.purchase/customerName = “CCL”) routeTo(r2) Potential conflict type: r1: if (xml.purchase/totalAmount 5000) routeTo(quickServers) r2: if (xml.purchase/totalAmount 20000) routeTo(superServers) Algorithm: Build tree with the same variable, check operator and value to see

33、if they are the same or lead to potential conflict, compare actions to decide conflict type or duplication. Editor can build these trees while a user enters rules and warns about conflict right away.,12/22/2000,Edward Chow ACSD Project Status 27,XML Tag Value Extraction,A xmlContentExtract() is buil

34、t to extract the tag values of a list of unique tag sequences. It is based on clark coopers expat 1.0 xmlparser. Its argument include the pointer to an XML document, the pointer to the array of strings (unique xml tag squences we follow the xsl selector syntax), and the number of sequences. It retur

35、n the list of a structure node, with the tag sequence, its attribute, and its value. Currently, it supports one attribute and tag sequece needs to be unique.,12/22/2000,Edward Chow ACSD Project Status 28,Status of UCCS ACSD Project,A Linux-based content switch prototype is almost complete. It is bas

36、ed on Linux-2.2.16-3 and lvs. ip_forward.c, ip_masq.c, ip_vs.c are modified to implement basic TCP delay binding. Preliminary tests had real server return web document and discovered the client retransmission problem. ip_cs.c are added for most of the content switching functions. http header extract

37、ion and xml content extract code are being integrated in for testing. A simple Java-based ruleEdit program was created for rule editing.,12/22/2000,Edward Chow ACSD Project Status 29,Related Load Balancing Research Results,Modified Apache status module to report Total bytes to be transferred by chil

38、d processes Average document transfer speed Modified LB-DNS to receive server status and bandwidth probing results. LB-DNS returns IP-address of the best server based a weight contributed by both server load and bandwidth. Modified WebStone benchmark to test the performance of load balancing web ser

39、ver clusters.,12/22/2000,Edward Chow ACSD Project Status 30,Load balancing Systems,Modified Web Server 1,Modified Web Server n,Statistics Gathering Daemon,LBA: Modified DNS,Server Delay,Request for Web pages,Server Ranking /tmp/StatFile,Bandwidth Probe Results,12/22/2000,Edward Chow ACSD Project Sta

40、tus 31,Connection Rate: LBA vs. Round-Robin,Round robin only run once,12/22/2000,Edward Chow ACSD Project Status 32,Conclusion,Content switch with generic rules can be easily configured for wide-variety of value-added services: Load balancing/High Available server farm. Premium services Firewall Ban

41、dwidth control/Traffic shaping Require efficient SW/HW architecture and rule matching algorithms to reduce processing overhead. Content rule design/conflict detection are important and challenging. TCP delay binding can be improved. Servicing multiple requests in keep alive session introduces interesting problem.,