Nahum E. et. al., 2002. Performance Issues in WWW servers. IEEE/ACM Transactions on Networking, Vol. 10, No. 1.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....method may work well for small classes. The application load time should be reduced, and the user should be able to interact with the application rather quickly. In practice, however, Web browsers make quite a few connections to retrieve class files. HTTP 1. 0, which is used by most Web servers [4], allows one request (e.g. for a class file) per connection. Therefore, if many class files are needed, many requests must be made, resulting in large communication overhead. The number of requests (thus, connections) made can be reduced by bundling and compressing class files into one file [5] ....

E. Nahum, T. Barzilai and D. D. Kandlur, "Performance Issues in WWW Servers," IEEE/ACM Transactions on Networking, vol. 10, no. 1, pp. 2-11.

....of inefficiencies in SILK that can affect performance. For example, SILK uses Scout Messages as its packet abstraction. This requires copying data between a Linux sk buff and a Message when moving a packet between SILK and Linux. Byte copying is one of the most expensive operations for Web servers [19]. The copy could be avoided by integrating Messages and sk buffs; this optimization is in progress. Also, path creation and destruction are fairly heavyweight operations. Table 1 shows some microbenchmarks for SILK averaged across runs for different file sizes. We note that caching SILK paths for ....

E. Nahum, T. Barzilai, and D. Kandlur. Performance issues in www servers. In Proc. of ACM SIGMETRICS, May 1999.

....Terms Algorithms, Management, Measurement, Performance, Experimentation. Keywords Web server, client perceived response time. 1. INTRODUCTION Appears in the Proceedings of the ACM SIGMETRICS 2002 The focus of web server performance is shifting from throughput and utilization benchmarks [22, 4, 24] to guaranteeing delay bounds for di#erent classes of clients [20, 34, 17, 28, 12, 2, 27, 9, 5] Providers of web services are faced with the challenge of providing di#erentiated services that guarantee bounds on client perceived response times while at the same time maximizing throughput. In ....

E. Nahum, T. Barzilai, and D. Kandlur. Performance issues in WWW servers. ACM SIGMETRICS

.... contrast to our study which investigates the tail of the load distribution) Finally, other issues such as strong and weak consistency mechanisms [6, 17, 32] cache replacement algorithms [5, 20] hierarchical proxy caching [29, 31, 33] prefetching techniques [14, 21, 25, 34] and protocol issues [3, 24] have also received significant research attention; these efforts are complementary to our current study. 6 Concluding Remarks In this paper, we examined the potential benefits of web proxy caches in improving the effective capacity of servers and networks. Since servers and networks are ....

E. Nahum, T. Barzilai, and D. Kandlur. Performance issues in WWW servers. In Proceedings of ACM SIGMERTICS'99 conference, Atlanta, GA, pages 216--217, May 1999.

....1 in a unicast environment, recent research efforts have # In this paper, we use the term Internet servers to refer to web servers, caches, proxies, reverse proxies, surrogates, etc. focused on operating system optimizations (e.g. memory subsystem, file system, network stack, etc. 29] 28] [25], 19] 18] Generally, these optimizations fall under two categories: 1) optimizations that improve resource allocation decisions, and (2) optimizations that boost resource utilization. Examples of approaches that aim to improve resource allocation decisions include the use of better cache ....

E. N. Nahum, Tsipora Barzilai, and Dilip Kandlur. Performance issues in WWW servers. In ACM Sigmetrics, 1999.

....using unicast based overlay networks [20] 14] Unicast based Techniques: To improve the scalability of Internet servers 1 in a unicast environment, recent research efforts have focused on operating system optimizations (e.g. memory subsystem, file system, network stack, etc. 29] 28] [25], 19] 18] Generally, these optimizations fall under two categories: 1) optimizations that improve resource allocation decisions, and (2) optimizations that boost resource utilization. Examples of approaches that aim to improve resource allocation decisions include the use of better cache ....

E. N. Nahum, Tsipora Barzilai, and Dilip Kandlur. Performance issues in WWW servers. In ACM Sigmetrics, 1999.

....using unicast based overlay networks [26] 23] Unicast based Techniques: To improve the scalability of Internet servers 2 in a unicast environment, recent research efforts have focused on operating system optimizations (e.g. memory subsystem, file system, network stack, etc. 37] 36] [33], 22] 21] Generally, these optimizations fall under two categories: 1) optimizations that improve resource allocation decisions, and (2) optimizations that boost resource utilization. Examples of approaches that aim to improve resource allocation decisions include the use of better cache ....

E. N. Nahum, Tsipora Barzilai, and Dilip Kandlur. Performance issues in WWW servers. Technical report, IBM, February 1999.

....interfaces or modify existing ones as necessary to improve performance. For example, the spread of the World Wide Web and the focus on the performance of Web servers has highlighted shortcomings of the UNIX I O system interfaces, where changes are notoriously hard to incorporate and standardize [Nahum99]. In another example, the continuous evolution of the internal Linux kernel interfaces does not allow interoperation with proprietary binary only device drivers. This is because changes to the interfaces cannot be propagated into the drivers without rebuilding them from source. One of the key ....

....equivalent system with a more integrated structure as measured by end to end performance. We implemented a simple, single threaded, event driven Web server application using a TCP stack component specialized for HTTP traffic. The specializations aimed to reduce the per connection latency (see [Nahum99]) We ran the Webstone benchmark to measure Web server performance on Pebble (with specialized TCP) and OpenBSD (with a standard TCP network stack) Pebble includes the following relevant components which run in separate protection domains: interrupt dispatcher, scheduler, ethernet driver, TCP ....

E. Nahum et al. Performance Issues in WWW Servers, IBM Technical Report, 1999

....architecture and one that employs a pool of threads, and evaluate their performance on UNIX systems as well as Windows NT using the WebStone benchmark. Various researchers have analyzed the processing costs of the different steps of HTTP request serving and have proposed improvements. Nahum et al. [25] compare existing high performance approaches with new socket APIs and evaluate their work on both single file tests and other benchmarks. Yiming Hu et al. 18] extensively analyze an earlier version of Apache and implement a number of optimizations, improving performance especially for smaller ....

E. Nahum, T. Barzilai, and D. Kandlur. Performance Issues in WWW Servers. submitted for publication.

....server without modifying the operating system. It caches files in user space with mmap( caches stat( information and URI lookups, and exploits writev( Flash has been shown to be significantly faster than Apache. We extended Flash in several ways to obtain the maximum performance in our system [20]. First, we modified it to poll( rather than select( Banga and Mogul [5] have shown that event driven servers that use select( such as Flash, can suffer performance problems when managing large numbers of active connections. To avoid these problems, we adapted Flash to use poll( a similar ....

Erich M. Nahum, Tsipora Barzilai, and Dilip Kandlur. Performance issues in WWW servers (extended abstract). In Proceedings of the ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, Atlanta, GA, May 1999. An extended version has been accepted to appear in IEEE/ACM Transactions on Networking.

....how our optimizations improve server throughput across requests for different file sizes as measured by WebStone. Table 2 shows the aggregate increase in server throughput as our optimizations are incrementally added as measured by SURGE. Interested readers should consult the IBM research report [6] for more details. We summarize our findings as follows: ffl new socket functions. We evaluate the proposed socket functions acceptex( and send file( We find little or no increase in performance using the acceptex( function, on either process based or thread based WWW servers. In addition, ....

Erich M. Nahum, Tsipora Barzilai, and Dilip Kandlur. Performance issues in WWW servers. IBM Research Report, May 1999.

No context found.

Nahum E. et. al., 2002. Performance Issues in WWW servers. IEEE/ACM Transactions on Networking, Vol. 10, No. 1.

No context found.

Nahum E. et. al., 2002. Performance Issues in WWW servers. IEEE/ACM Transactions on Networking, Vol. 10, No. 1.

No context found.

Nahum, E., Barzilai, T., and Kandlur, D., "Performance issues in WWW servers," IEEE/ACM Transactions on Networking, vol. 10, pp. 2--11, Feb. 2002.

No context found.

E. Nahum, T. Barailai, D. Kandlur, Performance Issues in WWW Servers, Proceedings of the international conference on Measurement and modeling of computer systems, 1999

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