V. N. Padmanabhan and J. C. Mogul. Improving HTTP Latency. In Proc. Second International WWW Conference, October 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....contributes most significantly to variability in total delay. I. INTRODUCTION Many aspects of current Internet infrastructure and protocols have been developed or enhanced as a direct result of careful analysis of network measurement data. Examples include application level protocols such as HTTP [1], 2] transport protocols such as TCP [3] 4] and TFRC [5] and distributed caching mechanisms such as [6] 7] While the body of measurement based work upon which these and other developments have been founded is significant, many of the complexities of Internet interactions (which if ....

V. Padmanabhan and J. Mogul, "Improving HTTP latency," Computer Networks and ISDN Systems, vol. 28, pp. 25--35, December 1995.

.... the document is present in an ISP (ICP queries [26] hashing function [20] routing [ In order to keep this delay low a caching hierarchy should not have more than three levels [6] 2) TCP delay, which is due to the slow start phase of the different TCP connections between every cache level [17]. The slow start is more relevant when the completion time of the document is small. The effect of this delay is reduced with persistent TCP connections [10] 3) Server delay, which is due to busy servers that need to deal with many requests for document updates from several national caches. 4) ....

V. N. Padmanabhan and J. Mogul, "Improving HTTP Latency", In Second World Wide Web Conference '94: Mosaic and the Web, pp. 995--1005, October 1994.

....to TCP connection latency and commences under slow start conditions. Additional load is placed upon the network, and the server has to service and maintain a large number of individual connections. Improvements to overcome the shortcomings in HTTP 1. 0 were suggested by Mogul and Padmanabhan [2]. Of particular relevance were persistent HTTP connections (P HTTP) and pipelining of multiple object requests over a single connection. Experimental data supported their recommendations and showed a significant reduction in overall latency for retrievals. A following paper by Mogul [3] ....

V. N. Padmanabhan and J. Mogul, "Improving HTTP latency," Computer Networks and ISDN Systems, vol. 28(1 -- 2), pp. 25 -- 35, 1995.

....querying. Even at the application level, complex applications like MultECommerce [26] specialize functionalities performed by nodes to achieve scalable systems. Third, clients can send multiple, and potentially very different, requests to the same server using a single persistent TCP connection [8, 23]. Persistent connections eliminate the overhead involved in setting up and tearing down a connection for each request. Unfortunately, they also make it hard to direct each request coming on the connection to the cluster node that is best suited to service it. In this paper, we present half pipe ....

....approach, content and services can be authored such that requests can be sent directly from the client to the appropriate servers. This approach has a drawback that it requires a client to open multiple connections with servers within a server cluster that wastes both server and network resources [13, 23]. Also, this approach exposes the configuration details of the server cluster to the content authoring process (and hence the clients) and prevents the cluster from achieving finegrained load balancing across servers. The second end to end approach uses a transport level connection migration ....

V. Padmanabhan and J. Mogul. Improving HTTP latency. In 2nd International WWW Conference, Chicago, IL, Oct 1994.

....study showed a larger average advertised receiver window size of 18KB [16] Small default values can limit a connection s throughput when the congestion window increases enough for the e ective limit on the sending rate to be the bu er sizes. This was shown to occur in several measurement studies [16, 23, 145]. However, some of these statements are based solely on the receive bu er size as advertised in the receiver window value, ignoring the e ect of the sender bu er size (perhaps because they know it is large enough not to be a factor) Situations where the send bu er size limits throughput have also ....

....the protocol in more details below. In HTTP 1.0 [30] each resource (i.e. object within a page) is transferred in a separate TCP connection, which is closed after the data is transferred . This creates a set of problems, which have been identi ed and addressed in the literature, notably in [131, 145], and described below. The rst problem relates to the management of TCP state at servers, where the succession of many short lived connections leaves the server with a large number of connections in the TIME WAIT state, which, according to the TCP standard, have to be kept for up to 4 minutes ....

[Article contains additional citation context not shown here]

Padmanabhan V., Mogul J., Improving HTTP Latency, in Computer Networks and ISDN Systems, December 1995.

....1 Scenario 2 Scenario 3 2.51 2.78 3.02 Table 3. Mean percentage of HO interrupted connections The observations from Table 3 are crucial since the HTTP uses TCP as a reliable transport protocol. Apart from the well known problems in the interaction of the stateless HTTP (ver.1. 0) with TCP [Pad95], Caceres and Iftode in [Cac95] have shown how TCP s congestion control mechanism undermines communication throughput during handovers. The above researchers have quantified the performance degradation in TCP connections caused by MT movement across cell boundaries. In an overlapping cell ....

V.N. Padmanabhan, and J.C. Mogul, "Improving HTTP Latency," Computer Networks and ISDN Systems Vol.28, 1995.

.... with the knowledge that most requests on the Web are for relatively small resources, means that reducing the number of connections that are necessary and holding them open (making them persistent) so that future requests can use them, has a significant positive e#ect on client performance [PM94, NGBS 97] In particular, for clients of forward proxies, time can be saved; the client can retain a persistent connection to the proxy, instead of taking the time to establish new connections with each origin server that a user might visit during a session. Use of persistent connections is ....

Venkata N. Padmanabhan and Je#rey C. Mogul. Improving HTTP latency. In Proceedings of the Second International World Wide Web Conference: Mosaic and the Web, pages 995--1005, Chicago, IL, October 1994.

....ance and control algorithms. However, while TCP does implement congestion control [17] many applications including the Web [5, 11] use several logically differ ent streams in parallel, using multiple concurrent TCP connections between the same pair of hosts. As several researchers have shown [2, 3, 27, 28, 42], these con current connections compete with rather than learn from each other about network conditions to the same receiver, and end up being unfair to other applications that use fewer connections. The ability to share conges tion information between concurrent flows is therefore a useful ....

....a general solution to these problems, application specific solutions have been proposed in the literature. Of particular importance are approaches that multiplex several logically distinct streams onto a single TCP connection at the application level, including Persistent connection HTTP (P HTTP [28], part of HTTP 1.1 [11] the Session Control Protocol (SCP) 39] and the MUX protocol [14] Unfortunately, these solutions suffer from two important drawbacks. First, because they are application specific, they require each class of applications (Web, real time streams, file transfers, etc. to ....

PADMANABHAN, V. N., AND MOGUL, J. C. Improving HTTP Latency. In Proc. Second International WWW Conference (Oct. 1994).

No context found.

V. N. Padmanabhan and J. C. Mogul. Improving HTTP Latency. In Proc. Second International World Wide Web Conference, October 1994. 15

No context found.

V. N. Padmanabhan and J. C. Mogul. Improving HTTP Latency. In Proc. Second International WWW Conference, October 1994.

No context found.

V. Padmanabhan and J. Mogul. Improving HTTP Latency. In Proc. Second International WWW Conference, Oct. 1994.

No context found.

Padmanabhan, V. N. and Mogul, J. 1995. Improving HTTP latency. Computer Networks and ISDN Systems 28, 25-35.

No context found.

V. N. Padmanabhan and J. Mogul, "Improving HTTP latency," in Proc. 2nd World Wide Web Conf. Mosaic and the Web, Chicago, IL, Oct. 1994, pp. 995--1005.

No context found.

Padmanabhan, Venkata N.; Mogul, Jerey C. Improving HTTP Latency. Proceedings Second WWW Conference 94: Mosaic and the Web. Chicago Illinois. Octuber 1994

No context found.

Venkata N. Padmanabhan and Jeffrey C. Mogul. Improving HTTP latency. Proceedings of the Second International World Wide Web Conference: Mosaic and the Web, October 1994.

No context found.

Padmanabhan, V. N., and Mogul, J. C. Improving HTTP Latency. In Proc. Second International WWW Conference (Oct. 1994).

No context found.

V. N. Padmanabhan and J. C. Mogul. Improving HTTP Latency. In Proc. Second International WWW Conference, October 1994.

No context found.

Venkata N. Padmanabhan and Jeffrey C. Mogul. Improving HTTP latency. In Second International World Wide Web Conference, Chicago, Illinois, USA, October 1994.

No context found.

Padmanabhan, V. N., and Mogul, J. C. Improving HTTP Latency. In Proc. Second International WWW Conference (Oct. 1994).

No context found.

) J. C. Mogul and V. N. Padmanabhan, "Improving HTTP Latency," Electronic Proceedings of the Second World Wide Web Conference '94: Mosaic and the Web, October 1994, http://www.ncsa.uiuc.edu/SDG/IT94/Proceedings/DDay/mogul/HTTPLatency.html.

No context found.

V. N. Padmanabhan and J. C. Mogul, "Improving HTTP Latency", Computer Networks and ISDN Systems, vol. 28, pp. 25-35, 1995

No context found.

Venkata N. Padmanabhan and Jeffrey C. Mogul. Improving HTTP latency. In Proceedings of the Second International World Wide Web Conference, October 1994.

No context found.

Venkata N. Padmanabhan and Jeffrey C. Mogul, "Improving HTTP latency," Computer Networks and ISDN Systems, vol. 28, no. 1--2, pp. 25--35, 1995.

No context found.

V. Padmanabhan and J. Mogul. Improving HTTP latency. In Proceedings of 2nd International World Wide Web Conference, Chicago, IL, October 1994.

No context found.

V. Padmanabhan, J. Mogul, Improving HTTP latency. Proc. nd International World Wide Web Conference, October 1994. Chicago, IL.

First 50 documents  Next 50

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