This study evaluates the performance benefits of using TCP Splice for handling non-cacheable Web objects and tunneling SSL connections in a Web proxy server. For this type of traffic, the Web proxy cannot draw any future benefits, yet it spends substantial CPU cycles moving data between server and client connections with no processing other than header manipulation. Previous studies showed the benefits of TCP Splice in reducing the overheads in proxy servers that do not need to access the content exchanged by client and server after the connection to server is established (e.g., firewalls). Our study is focused on a different type of proxy; for many of its requests, a Web proxy needs access the HTTP header sent by the server in order to assess the content cacheability. Our experimental results demonstrate that TCP Splice can benefit Web proxy servers, as well. TCP Splice enables 30-55\% reductions in per-request proxy overheads, while response time is reduced by up to 25\% for files larger than 10KBytes. Larger improvements are observed when handling SSL connections, in particular for small file sizes. Our evaluation is performed with a socket-level implementation of TCP Splice. Different from previously proposed TCP splice implementations, socket-level splicing allows splicing TCP connections with different characteristics. Moreover, experiments show that this implementation helps reduce the number of retransmission timeouts observed by the Web server, thus improving on the response time observed by the client. The experimental testbed includes an emulated WAN environment and benchmark applications for HTTP/1.0 Web clients, server, and proxy running on AIX machines.
By: Marcel Rosu, Daniela Rosu
Published in: RC22159 in 2001
LIMITED DISTRIBUTION NOTICE:
This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.
Questions about this service can be mailed to reports@us.ibm.com .