Date: Wed, 17 Apr 1996 13:12:53 -0400 From: dennis@etinc.com (dennis) To: Michael Smith <msmith@atrad.adelaide.edu.au> Cc: hackers@freebsd.org Subject: Re: TCP Window question Message-ID: <199604171712.NAA15988@etinc.com>
next in thread | raw e-mail | index | archive | help
>dennis stands accused of saying: >> >> Now for the more important question: Does anyone care to hazzard a guess as >> to the pct of "broken" implementations that will reject (or choke on) "out >> of sequence" packets? > >*laugh* That depends on the market you're looking at. If you're referring >to Unix stacks, almost zero, and you'd have to go back a long way to find >them. > >The out-of-order reassembly handling is necessary to handle lost or >corrupted packets. Any system that choked on such a circumstance would >do so as soon as an ethernet collision killed a passing TCP fragment. This seems obvious, but I'm wondering what all the ruckus is about as there is a lot of complaining about certain vendors "load balancing" techniques that send packets out of sequence fairly regularly. It seems that theres more than 1 implementation out there that will discard out of sequence packets thus requiring a retrans. It doesnt make too much difference, as the goal of any good technique would be to minimize the occurrance...but Id like to find a broken one to test with.... Dennis ---------------------------------------------------------------------------- Emerging Technologies, Inc. http://www.etinc.com Synchronous Communications Cards and Routers For Discriminating Tastes. 56k to T1 and beyond. Frame Relay, PPP, HDLC, and X.25 for BSD/OS, FreeBSD and LINUX
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199604171712.NAA15988>