H. Nielsen, Jim Gettys, Anselm Baird-Smith, et al. "Network Performance Effects of HTTP/1.1, CSS1, and PNG". June 42, 1997. Available from "http://www.w3.org/TR/NOTE-pipelining".  Home/Search   Document Not in Database   Summary   Related Articles   Check

....both the base page and subsequent embedded objects. Latency is thus reduced since with a single connection for everything there is only one handshake procedure and one slow start stage of TCP. There has been extensive studies on the performance of different versions of HTTP, see for example [25] [26]. Here we compare HTTP 1.0 when proxy is enabled vs. HTTP 1.1 when proxy is disabled. The connection setup of this part of our measurements corresponds to Figures 3(e) and 3(f) i.e. the connections do not go through the cache and that the connection is either end to end (proxy disabled) or ....

H. F. Nielsen et al., "Network Performance Effects of HTTP/1.1, Css1, and PNG," Tech. Rep., 1997, http://www.w3.org/Protocols/HTTP/Performance/Pipeline.html.

....over the same TCP connection, amortizing the TCP set up overhead. Pipelining lets the client send multiple requests over the same connection without waiting for responses. The server will send a stream of responses back. These features have been shown to reduce client latency and network traffic [8]. However, they do not eliminate all overheads of TCP, and in fact may introduce new performance penalties, especially when the bottleneck is at the server [9] Persistent connections increase the number of open connections at the server, which can have a significant negative effect on server ....

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley, "Network performance effects of HTTP/1.1, CSS1, and PNG," in Proceedings of ACM SIGCOMM'97 Conference, Sept. 1997.

....be maintained indefinitely, a decision on when a persistent connection is to be closed needs to be made. A fixed timeout period after which a server automatically closes a connection is suggested [1, 2] A quasi dynamic solution, where two fixed timers are used, is proposed in [2] Nielsen et al. [3] suggest a connection to be closed after servicing a fixed number of requests. Version 1.3.1 of the Apache Web server uses a combination of the solutions proposed in [2, 3] None of these solutions accounts for the variability of the server load during different hours of the day and different ....

....a connection is suggested [1, 2] A quasi dynamic solution, where two fixed timers are used, is proposed in [2] Nielsen et al. 3] suggest a connection to be closed after servicing a fixed number of requests. Version 1.3. 1 of the Apache Web server uses a combination of the solutions proposed in [2, 3]. None of these solutions accounts for the variability of the server load during different hours of the day and different days of the week. Adaptability is important since Web servers usually have periods of high load and periods of low load. We propose a solution that uses the inter arrival ....

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley, "Network performance effects of HTTP/1.1, CSS1, and PNG," in ACM SIGCOMM conference, pp. 155--166, Oct. 1997.

....is only one of the necessary steps for understanding the changes occurring in Web traffic. Research efforts on Web client workloads (e.g. 5] Web proxy workloads (e.g. 2] 7] 8] 15] 17] 18] 22] 27] 33] network traffic characterizations ( e.g. 34] as well as HTTP analyses (e.g. 4][21][23] 30] are all required in order better understand the Web. The remainder of the paper is organized as follows. Section 2 provides background information on the 1998 World Cup, focusing on the structure of the tournament. Section 3 introduces the World Cup Web site and describes the technology ....

H. Frystyk-Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. Wium-Lie and C. Lilley, "Network Performance Effects of HTTP/1.1, CSS1 and PNG", Proceedings of ACM SIGCOMM `97, Cannes, France, September 1997.

....and data compression tailored to web transmission. Although some previous work has accelerated web access through lossy image compression [7] 12] 25] this, by necessity, results some reduction in image quality. Transport protocol modifications designed to reduce the total web page delivery time [17][2] 19] also improve web performance, but further gains can be achieved by combining image coding and networking techniques as described in this paper. It is important to view web browsing as a form of remote display, similar to the X Windowing System [24] Browsing is not the bulk transfer of ....

....HTTP 1. 1 also supports compression of the HTML objects via the deflate coding of the public domain zlib compression library [9] This is a hybrid LZ77 [28] Huffman coding [10] When applied to HTML data, it can typically result in a greater than 3x reduction in the size of the HTML file [17]. TCP sessions [2] 18] have been proposed to allow sharing of state between related TCP connections, such as those connecting the same host pair, at the transport layer. If any connection in the session experiences congestion, all connections can reduce their windows. In this way the ....

Nielsen, Henrik Frystky, et al., "Network Performance Effects of HTTP/1.1, CSS1, and PNG." W3C NOTEpipelining -970624. June 24, 1997. http://www.w3c.org/Protocols/HTTP/Performance/ Pipeline.html.

....under overload. Mosberger and Jin describe httperf, a tool that attempts to provide a comprehensive benchmark for Web server measurement [Mosberger and Jin 1998] httperf supports HTTP 1. 1 style persistent connections, request pipelining, and chunked transfer encoding [Fielding et al. 1997; Nielsen et al. 1997]. An important concern in the design of this tool is the complete separation of the issues related to the actual generation of HTTP calls from those related to the specific workload and measurements that should be used. httperf uses a specially constructed method, which is similar to and partly ....

Nielsen, H., J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley (1997), "Network performance effects of HTTP/1.1, CSS1, and PNG," In Proceedings of the ACM SIGCOMM '97 Conference, Cannes, France, pp. 155--166.

....workload characteristics (as summarized in Table 1) ffl Persistent connections and pipelining: Studies of HTTP 1. 0 traffic have shown that individual HTTP transfers are typically short lived, making the round trip time for TCP connection establishment a significant portion of the total latency [3, 4]. In HTTP 1.1, HTTP connections are persistent by default, to avoid the round trip delays, as well as the processing and bandwidth overheads of repeatedly opening and closing connections. The move to persistent connections should result in appreciable differences in the arrival times of requests ....

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley, "Network performance effects of HTTP/1.1, CSS1, and PNG," in Proc. ACM SIGCOMM, pp. 155--166, August 1997. http://www.inria.fr/rodeo/sigcomm97/program.html.

....Read Request Parse Request Log Request DLL Cache Strategy . Figure 7: The Service Configurator Pattern in JAWS 2. 2 JAWS Web Server Performance Our research [20, 21] demonstrates that it is possible to improve server performance through superior server design (a similar observation was made in [22]) Thus, while a hardcoded server, i.e. one that uses fixed concurrency, I O, and caching strategies, can provide excellent performance, a flexible server framewor like JAWS need necessarily not perform poorly. Figure 8 below illustrates how the flexible nature of the JAWS framework enables it ....

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley, "Network Performance Effects of HTTP/1.1, CSS1, and PNG," in To appear in Proceedings of ACM SIGCOMM '97, 1997.

....each object that appears in the page, therefore, there is some more control traffic devoted to establishing connections. HTTP 1. 1, by means of persistent connections, tries to minimize the extra traffic generated in the establishment of each connection by having just one TCP connection per page [Nielsen 97] 3 All the protocols with which we carried out the trials use TCP as the transport protocol. 3. Trials The trials were carried out among two PCs running Linux OS. The characteristics of the PCs used for the trials were: Pccba4: Pentium MMX 166 RAM 80 Mb Ethernet card: 3Com 3c509 OS: Linux ....

Nielsen HF, Gettys J, Baird-Smith A et al. `Network Performance Effects of HTTP/1.1, CSS1, and PNG.' Proceedings of ACM SIGCOMM'97, Cannes (France), pp. 155-166, September 1997.

.... simply Not Modified responses to validate cached resources [9, 10] To avoid the latency and overhead of establishing a TCP connection for each Web transfer, extensions to HTTP permit proxies to maintain persistent connections to servers, which enable pipelining of multiple requests and responses [11, 12]. For example, a persistent connection permits embedded images in an HTML document to be downloaded without new TCP connections. To reduce the number of TCP connections, a proxy can allow multiple clients to share a single persistent connection to a server. Similarly, a client via a persistent ....

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley, "Network performance effects of HTTP/1.1, CSS1, and PNG," in Proc. ACM SIGCOMM, pp. 155--166, August 1997. http://www.inria.fr/rodeo/sigcomm97/ program.html.

....Clients Thread Request Thread Pool TransmitFile Figure 23: Experiment Results from 5M File 4. 4 A Summary of Techniques for Optimizing Web Servers From our research, we have found that it is possible to improve server performance with a superior server design (a similar observation was made in [34]) Thus, while it is undeniable that a hard coded server (i.e. one that uses fixed concurrency, I O, and caching strategies) can provide excellent performance, a flexible server framework, such as JAWS, does not necessarily correlate with poor performance. This section summarizes the most ....

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley, "Network Performance Effects of HTTP/1.1, CSS1, and PNG," in To appear in Proceedings of ACM SIGCOMM '97, 1997.

....wait to receive the response for one request before sending another request on the same connection. In fact, a client could send an arbitrarily large number of requests over a TCP connection before receiving any of the responses. This practice, known as pipelining, can greatly improve performance [22] . It avoids the need to wait for network round trips, and it makes the best possible use of the TCP protocol. 6 Message transmission HTTP messages may carry a body of arbitrary length. The recipient of a message needs to know where the message ends. The sender can use the Content Length header, ....

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley, "Network Performance Effects of HTTP/1.1, CSS1, and PNG," in Proc. SIGCOMM '97, (Cannes, France), September 1997.

....implemented persistent and pipelined client connections, but as of August 1997 have not yet implemented persistent connections to servers or neighbor caches. The World Wide Web Consortium has reported observing both reduced latency and reduced number of packets transmitted with pipelined requests[9]. 2.3.5 FTP requests handled internally We originally believed that the FTP protocol was too complex to implement inside the cache itself and so used an external module, ftpget , which was ugly and difficult to understand. Squid 1.2 no longer uses a separate process, the FTP functionality is ....

H. F. Nielsen et al., "Network performance effects of http/1.1, css1, and png," June 1997. http://www. w3.org/pub/WWW/Protocols/HTTP/Performance/Pipeline.html.

....a client to make multiple requests of the same server in a short amount of time. Analysis of these performance problems and results from a prototype implementation are available [26] 30] Implementation experience and measurements of actual HTTP 1. 1 (RFC 2068) implementations show good results [39]. Alternatives have also been explored, for example, T TCP [27] Persistent HTTP connections have a number of advantages: By opening and closing fewer TCP connections, CPU time is saved in routers and hosts (clients, servers, proxies, gateways, tunnels, or caches) and memory used for TCP ....

Nielsen, H.F., Gettys, J., Baird-Smith, A., Prud'hommeaux, E., Lie, H., and C. Lilley. "Network Performance Effects of HTTP/1.1, CSS1, and PNG," Proceedings of ACM SIGCOMM '97, Cannes France, September 1997.[jg656]

....multiple requests of the same server in a short amount of time. Analyses of these performance problems are available [30] analysis and results from a prototype implementation are in [26] Implementation experience and measurements of actual HTTP 1. 1 (RFC 2068) implementations show good results [39]. Alternatives have also been explored, for example, T TCP [27] Persistent HTTP connections have a number of advantages: By opening and closing fewer TCP connections, CPU time is saved, and memory used for TCP protocol control blocks is also saved. HTTP requests and responses can be ....

Nielsen, H.F., Gettys, J., Baird-Smith, A., Prud'hommeaux, E., Lie, H., and C. Lilley. "Network Performance Effects of HTTP/1.1, CSS1, and PNG," Proceedings of ACM SIGCOMM '97, Cannes France, September 1997.

....a client to make multiple requests of the same server in a short amount of time. Analysis of these performance problems and results from a prototype implementation are available [26] 30] Implementation experience and measurements of actual HTTP 1. 1 (RFC 2068) implementations show good results [39]. Alternatives have also been explored, for example, T TCP [27] Persistent HTTP connections have a number of advantages: By opening and closing fewer TCP connections, CPU time is saved in routers and hosts (clients, servers, proxies, gateways, tunnels, or caches) and memory used for TCP ....

Nielsen, H.F., Gettys, J., Baird-Smith, A., Prud'hommeaux, E., Lie, H., and C. Lilley. "Network Performance Effects of HTTP/1.1, CSS1, and PNG," Proceedings of ACM SIGCOMM '97, Cannes France, September 1997.[jg592]

....requests of the same server in a short amount of time. Analyses of these performance problems are available [30] 27] analysis and results from a prototype implementation are in [26] Implementation experience and measurements of actual HTTP 1. 1 (RFC 2068) implementations show good results [39]. Alternatives have also been explored, for example, T TCP [27] Persistent HTTP connections have a number of advantages: By opening and closing fewer TCP connections, CPU time is saved, and memory used for TCP protocol control blocks is also saved. HTTP requests and responses can be ....

Nielsen, H.F., Gettys, J., Baird-Smith, A., Prud'hommeaux, E., Lie, H., and C. Lilley. "Network Performance Effects of HTTP/1.1, CSS1, and PNG," Proceedings of ACM SIGCOMM '97, Cannes France, September 1997.

....a client to make multiple requests of the same server in a short amount of time. Analysis of these performance problems and results from a prototype implementation are available [26] 30] Implementation experience and measurements of actual HTTP 1. 1 (RFC 2068) implementations show good results [39]. Alternatives have also been explored, for example, T TCP [27] Persistent HTTP connections have a number of advantages: INTERNET DRAFT HTTP 1.1 November 18, 1998 Fielding, et al. [Page 31] By opening and closing fewer TCP connections, CPU time is saved in routers and hosts (clients, servers, ....

Nielsen, H.F., Gettys, J., Baird-Smith, A., Prud'hommeaux, E., Lie, H., and C. Lilley. "Network Performance Effects of HTTP/1.1, CSS1, and PNG," Proceedings of ACM SIGCOMM '97, Cannes France, September 1997.[jg642]

....multiple requests of the same server in a short amount of time. Analyses of these performance problems are available [30] analysis and results from a prototype implementation are in [26] Implementation experience and measurements of actual HTTP 1. 1 (RFC 2068) implementations show good results [39]. Alternatives have also been explored, for example, T TCP [27] Persistent HTTP connections have a number of advantages: By opening and closing fewer TCP connections, CPU time is saved, and memory used for TCP protocol control blocks is also saved. HTTP requests and responses can be ....

Nielsen, H.F., Gettys, J., Baird-Smith, A., Prud'hommeaux, E., Lie, H., and C. Lilley. "Network Performance Effects of HTTP/1.1, CSS1, and PNG," Proceedings of ACM SIGCOMM '97, Cannes France, September 1997.[jg243]

No context found.

H. Nielsen, Jim Gettys, Anselm Baird-Smith, et al. "Network Performance Effects of HTTP/1.1, CSS1, and PNG". June 42, 1997. Available from "http://www.w3.org/TR/NOTE-pipelining".

No context found.

H. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. Lie, and C. Lilley, "Network Performance Effects of HTTP/1.1, CSS1, and PNG", work in progress.

No context found.

H. Nielsen, Jim Gettys, Anselm BairdSmith, et al. "Network Performance Effects of HTTP/1.1, CSS1, and PNG". June 42, 1997. Available from http://www.w3.org/TR/NOTEpipelining.

No context found.

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'Hommeaux, H. W. Lie, C. Lilley, "Network Performance Effects of HTTP/1.1, CSS1 and PNG", Proc. of ACM SIGCOMM'97, Sep. 1997.

No context found.

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, C. Lilley, "Network performance effects of HTTP/1.1 CSS1, and PNG," available as http://www/w3/org/TR/note-pipelining.

No context found.

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, and C. Lilley, "Network performance effects of HTTP/1.1, CSS1, and PNG," in Proceedings of ACM SIGCOMM, (Cannes, France), pp. 155--166, August 1997. http://www.inria.fr/rodeo/sigcomm97/program.html.

No context found.

H. F. Nielsen, J. Gettys, A. Baird-Smith, E. Prud'hommeaux, H. W. Lie, C. Lilley, "Network performance effects of HTTP/1.1 CSS1, and PNG," available as http://www/w3/org/TR/note-pipelining.

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