M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived web performance. In Proc. of the Sixth Int. Workshop on Web Caching and Content Distribution, June 2001.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....dynamically generated web pages might contain embedded objects served by the CDN. Previous research has investigated the use and effectiveness of content delivery networks [14] although the proprietary and closed nature of these systems tends to impede investigation. Two recent studies [22, 23] confirm that CDNs reduce average download response times, but that DNS redirection techniques add noticeable overhead because of DNS latencies. In another study [18] the authors argue that the true benefit of CDNs is that they help clients avoid the worst case of badly performing replicas, ....

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived web performance. In Proc. of the Sixth Int. Workshop on Web Caching and Content Distribution, June 2001.

....although dynamically generated web pages might contain embedded objects served by the CDN. Previous research has investigated the use and effectiveness of content delivery networks [14] although the proprietary and closed nature of these systems tends to impede investigation. Two recent studies [22, 23] confirm that CDNs reduce average download response times, but that DNS redirection techniques add noticeable overhead because of DNS latencies. In another study [18] the authors argue that the true benefit of CDNs is that they help clients avoid the worst case of badly performing replicas, ....

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived web performance. In Proc. of the Sixth Int. Workshop on Web Caching and Content Distribution, June 2001.

....only one small object, and there was no comparison with downloading from the origin server. A study done in the context of developing the request mirroring Medusa Web proxy, evaluated the performance of one CDN (Akamai) by downloading the same objects from CDN servers and origin servers [37]. The study was done only for a single user workload, but showed significant performance improvement for those objects that were served by the CDN, when compared with the origin server. More recently, Krishnamurthy et al. studied the performance of a number of commercial CDNs from the vantage ....

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived web performance. In Proceedings of International Web Caching and Content Delivery Workshop (WCW), Boston, MA, June 2001. Elsevier.

....Surge [5] and Web Polygraph [30] Some use artificial traces; some base their workloads on data from realworld traces. They are not, however, principally designed to use a live workload or live network connection, and are generally incapable of handling prefetching proxies. Koletsou and Voelker [19] built the Medusa Proxy, which is designed to measure user perceived Web performance. It operates similarly to the Multiplier in SPE in that it duplicates requests to different Web delivery systems and compares results. It also can transform requests, e.g. from Akamaized URLs to URLs to the ....

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived Web performance. In Web Caching and Content Delivery: Proceedings of the Sixth International Web Content Caching and Distribution Workshop (WCW'01), Boston, MA, June 2001.

....latency costs. Both of these increase the number of hops through which content must travel, directly increasing the cost of a Web retrieval. While access to some objects served by CDNs are getting better, these still represent only a relatively small fraction of a typical user s content [KV01, BV02] Bandwidth demands continue to increase. New users are still getting connected to the Internet in large numbers. Even as growth in the user base slows, demand for increased bandwidth will continue as high bandwidth media such as audio and video increase in popularity. If the price is ....

....and other content networking equipment. It includes highperformance HTTP clients and servers to generate artificial Web workloads with realistic characteristics. Web Polygraph has been used to benchmark proxy cache performances in multiple Web cache o#s [RW00, RWW01] Koletsou and Voelker [KV01] built the Medusa Proxy, which is designed to measure user perceived Web performance. It operates similarly to our Multiplier in that it duplicates requests to di#erent Web delivery systems and compares results. It also can transform requests, e.g. from Akamaized URLs to URLs to the customer s ....

Mimika Koletsou and Geo#rey M. Voelker. The Medusa proxy: A tool for exploring user-perceived Web performance. In Web Caching and Content Delivery: Proceedings of the Sixth International Web Content Caching and Distribution Workshop (WCW'01), Boston, MA, June 2001.

....of caching proxies and other content networking equipment. It includes high performance HTTP clients and servers to generate artificial Web workloads with realistic characteristics. Web Polygraph has been used to benchmark proxy cache performances in Web cache o#s [32] Koletsou and Voelker [26] built the Medusa Proxy, which is designed to measure user perceived Web performance. It operates similarly to our Multiplier, in that it duplicates requests to di#erent Web delivery systems and compares results. It also can transform requests, e.g. from Akamaized URLs to URLs to the customer s ....

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived Web performance. In Proceedings of the Sixth International Workshop on Web Caching and Content Distribution (WCW'01), Boston, MA, June 2001.

....traces using an instrumented client; this could capture every user reference. Instrumented browsers have been used to collect traces from small user populations [14, 16] It is difficult to instrument popular browsers today because source code is unavailable, but a client proxy such as Medusa [29] can collect much of the same data. The main problem with client end data collection is the difficulty of deployment across a large client sample. In collecting the anonymized traces we analyze, WebTV employed a non caching, cache busting proxy. A cache busting proxy marks as uncachable all of ....

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived Web performance. In Proc. 6th Web Caching Workshop, June 2001.

....traces using an instrumented client; this could capture every user reference. Instrumented browsers have been used to collect traces from small user populations [14, 16] It is difficult to instrument popular browsers today because source code is unavailable, but a client proxy such as Medusa [29] can collect much of the same data. The main problem with client end data collection is the difficulty of deployment across a large client sample. In collecting the anonymized traces we analyze, WebTV employed a non caching, cache busting proxy. A cache busting proxy marks as uncachable all of ....

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived Web performance. In Proc. 6th Web Caching Workshop, June 2001.

....a single 3 4KB image they found that the CDNs appeared to use the DNS mechanisms not necessarily to select optimal servers, but to avoid selection of bad servers, though it is hard to know how to generalize their study given its limited scope. In more recent work, the mirroring proxy Medusa [9] examines a single CDN (Akamai) in a technique similar to [3] The study presents performance improvement results from a single user point of view for a small workload. DNS timeout effects are ignored in their study due to a fairly small inter request interval. On the Use and Performance of ....

....origin sites from images being served by those sites. In comparing CDN performance with that of origin servers a seemingly ideal test would retrieve two versions of an origin server s page one using a CDN to serve content and the other not. This approach was used for two previous studies [3, 9] . However, a straightforward application of this test for a origin server employing a CDN for partial site DNS redirection causes content to be retrieved from an origin server that it expects to be retrieved from the CDN. Even if the origin server contains the content it is unlikely to be ....

M. Koletsou and G. M. Voelker, "The medusa proxy: A tool for exploring user-perceived web performance," in Proceedings of the 6th International Web Caching Workshop and Content Delivery Workshop, (Boston, MA), June 2001. http://www-cse.ucsd.edu/~voelker/pubs/medusa-wcw01.pdf.

....a single 3 4KB image they found that the CDNs appeared to use the DNS mechanisms not necessarily to select optimal servers, but to avoid selection of bad servers, though it is hard to know how to generalize their study given its limited scope. In more recent work, the mirroring proxy Medusa [9] examines a single CDN (Akamai) in a technique similar to [3] The study presents performance improvement results from a single user point of view for a small workload. DNS timeout effects are ignored in their study due to a fairly small interrequest interval. Our study evaluates response time ....

....origin sites from images being served by those sites. In comparing CDN performance with that of origin servers a seemingly ideal test would retrieve two versions of an origin server s page one using a CDN to serve content and the other not. This approach was used for two previous studies [3] [9]. However, a straightforward application of this test for a origin server employing a CDN for partialsite DNS redirection causes content to be retrieved from an origin server that it expects to be retrieved from the CDN. Even if the origin server contains the content it is unlikely to be optimized ....

M. Koletsou and G. M. Voelker, "The medusa proxy: A tool for exploring user-perceived web performance," in Proceedings of the 6th International Web Caching Workshop and Content Delivery Workshop, (Boston, MA), June 2001. http://www-cse.ucsd.edu/voelker/ pubs/medusa-wcw01.pdf.

No context found.

M. Koletsou and G. M. Voelker. The medusa proxy: A tool for exploring user-perceived web performance. In Sixth International Workshop on Web Caching and Content Distribution, Boston, MA 2001.

....Web performance of downloading entire pages, and how various common Web enhancements impact overall page performance. Extensive previous work has studied how various techniques, such caching (e.g. 5, 7,12,16,19] prefetching (e.g. 1, 3,4,14] content distribution networks (CDNs) e.g. [6, 8, 10, 11]) and DNS resolution (e.g. 9, 17, 18] impact the performance of downloading individual objects. But when browsing the Web, users are much more concerned with the performance of entire pages. Although whole page performance is determined by the performance of its components, optimizations like ....

....up a number of interesting questions about userperceived Web performance. How does whole page performance compare to individual object performance How do various downloading optimizations improve whole page performance Content distribution networks (CDNs) can improve object download performance [8, 10, 11], but how do CDNs impact whole page latency given that typically only a subset of objects comprising a page are fetched from a CDN Since DNS resolution can increase download latency [2, 18] particularly when using CDNs [11] how does DNS caching amortize resolution costs across all of the ....

[Article contains additional citation context not shown here]

M. Koletsou and G. M. Voelker. The medusa proxy: A tool for exploring user-perceived web performance. In Sixth International Workshop on Web Caching and Content Distribution, Boston, MA 2001.

No context found.

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived web performance. In Proc. of the Sixth Int. Workshop on Web Caching and Content Distribution, June 2001.

No context found.

M. Koletsou and G. Voelker. The medusa proxy: A tool for exploring user-perceived web performance. In 6th International Web Caching Workshop and Content Delivery Workshop, Boston, MA, June 2001.

No context found.

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived web performance. In Proc. of the Sixth Int. Workshop on Web Caching and Content Distribution, June 2001.

No context found.

M. Koletsou and G. Voelker, "The Medusa Proxy: A Tool for Exploring User-Perceived Web Performance," Proc. Sixth Int'l Web Caching Workshop and Content Delivery Workshop, June 2001.

No context found.

M. Koletsou and G. Voelker. The medusa proxy: A tool for exploring user-perceived web performance. In 6th International Web Caching Workshop and Content Delivery Workshop, Boston, MA, June 2001.

No context found.

M. Koletsou and G. M. Voelker. The Medusa proxy: A tool for exploring user-perceived web performance. In Proc. of the Sixth Int. Workshop on Web Caching and Content Distribution, June 2001.

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