Return-Path: Received: by vger.rutgers.edu via listexpand id <157671-8093>; Sun, 17 Jan 1999 22:30:07 -0500 Received: by vger.rutgers.edu id <157661-8100>; Sun, 17 Jan 1999 22:29:56 -0500 Received: from [204.193.135.230] ([204.193.135.230]:12395 "EHLO linux42.dn.net" ident: "NO-IDENT-SERVICE[2]") by vger.rutgers.edu with ESMTP id <157651-8100>; Sun, 17 Jan 1999 22:29:44 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <13986.43678.718410.351098@wander.auton.org> Date: Sun, 17 Jan 1999 22:29:34 -0500 (EST) To: linux-kernel@vger.rutgers.edu From: lk@winux.com Subject: TCP Splice alleviates proxy bottleneck X-Mailer: VM 6.62 under 21.2 "Aglaia" XEmacs Lucid (beta3) Sender: owner-linux-kernel@vger.rutgers.edu Content-Length: 1940 Lines: 42 Anyone who's interested in high performance network development on Linux should take a look at what these folks are doing. It seems to be a very clever way to eliminate the drudge work that an application level proxy performs without significant impact on the application code. Looks like a great performance booster. Very nice. The paper says that this is all running on BSDI. Wish they'd picked Linux.. :-) Application Layer Proxy Performance Using TCP Splice. David Maltz, Pravin Bhagwat. IBM Technical Report RC-21139, March 1998. (submitted for publication). TCP Splicing for Application Layer Proxy Performance David Maltz, Pravin Bhagwat Application layer proxies already play an important role in today's networks, serving as firewalls and HTTP caches --- and their role is being expanded to include encryption, compression, and mobility support services. Current application layer proxies suffer major performance penalties as they spend most of their time moving data back and forth between connections; context switching and crossing protection boundaries for each chunk of data they handle. We present a technique called TCP Splice that provides kernel support for data relaying operations which runs at near router speeds. In our lab testing, we find SOCKS firewalls using TCP Splice can sustain a data throughput twice that of normal firewalls, with an average packet forwarding latency 30 times less. http://www.cs.umd.edu/users/pravin/TR-21139.ps.gz Here are some other publications by one of the authors: http://www.cs.umd.edu/users/pravin/publist.htm The MSOCKS paper also discusses TCP Splice. -lda - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/