Gentoo Archives: gentoo-performance

From: Ritesh Kumar <ritesh@××××××.edu>
To: gentoo-performance@l.g.o
Subject: Re: [gentoo-performance] TCP perfomance
Date: Sat, 07 Jun 2008 16:54:44
Message-Id: f47983b00806070954y7105f548i71356ad459688f28@mail.gmail.com
In Reply to: Re: [gentoo-performance] TCP perfomance by Kevin Faulkner
1 There are quite a few things that make TCP performance degrade. Is there a
2 specific reason why you thought about TCP timestamps?
3 A good way to know is to launch a TCP connection and get it to a stage where
4 you can clearly see its performance suffer. Then first take a look at the
5 CPU usage... if its 100% only then you should conclude that the problem is
6 with the CPU usage aspect of it.
7 In my experience, there can be many other things that may go wrong. There
8 are quite a few 'TCP tuning' faqs around the Internet. You might want to go
9 through a few to see some commonly used solutions.
10
11 OTOH, TCP timestamps are an important part of TCP... may be not for a LAN
12 like system but definitely for a complicated network like the Internet,
13 where timestamps help TCP to easily infer round trip times with much more
14 precision than otherwise.
15
16 _r
17
18 On Fri, Jun 6, 2008 at 1:26 PM, Kevin Faulkner <kevlar.kernel@×××××.com>
19 wrote:
20
21 > Miguel Sousa Filipe wrote:
22 >
23 >> There is no such thing has a TCP timestamp:
24 >> http://freebie.fatpipe.org/~mjb/Drawings/TCP_Header.png<http://freebie.fatpipe.org/%7Emjb/Drawings/TCP_Header.png>
25 >>
26 >> so, that doesn't make any sense...
27 >>
28 > Your right and I'm wrong.
29 > Its not in the header, its thrown on at the end....
30 > try doing
31 > cat /proc/sys/net/ipv4/tcp_timestamps
32 >
33 > http://www.networksorcery.com/enp/protocol/tcp/option008.htm
34 > http://www.securiteam.com/securitynews/5NP0C153PI.html
35 > --
36 > gentoo-performance@l.g.o mailing list
37 >
38 >