| 1 |
Miguel Sousa Filipe wrote: |
| 2 |
> On Fri, Jun 6, 2008 at 6:26 PM, Kevin Faulkner <kevlar.kernel@×××××.com> wrote: |
| 3 |
>> Miguel Sousa Filipe wrote: |
| 4 |
>>> There is no such thing has a TCP timestamp: |
| 5 |
>>> http://freebie.fatpipe.org/~mjb/Drawings/TCP_Header.png |
| 6 |
>>> |
| 7 |
>>> so, that doesn't make any sense... |
| 8 |
>> Your right and I'm wrong. |
| 9 |
>> Its not in the header, its thrown on at the end.... |
| 10 |
> |
| 11 |
> ...at the end of the tcp header, it's a tcp option. (I didn't |
| 12 |
> understood that the first time I read this email). |
| 13 |
> It might be good to disable all tcp optional headers.. |
| 14 |
> Also there might be important to look at issues such has: |
| 15 |
> - mtu size |
| 16 |
> - tcp window |
| 17 |
> - set the don't fragment flag (this can offload the routers and |
| 18 |
> optimize the mtu for the whole connection path) |
| 19 |
> - use the BIC algorithm (from what I've read, that's my default choice |
| 20 |
> nowadays... but for some specific workload there might be better |
| 21 |
> algorithms) |
| 22 |
|
| 23 |
I haven't played around with different congestion algorithms all that |
| 24 |
much, I generally have stuck with Westwood. MTU size... with all the |
| 25 |
different machines we have, I prefer to stay with the default. Window |
| 26 |
size is a good point. |
| 27 |
|
| 28 |
>> try doing |
| 29 |
>> cat /proc/sys/net/ipv4/tcp_timestamps |
| 30 |
>> |
| 31 |
>> http://www.networksorcery.com/enp/protocol/tcp/option008.htm |
| 32 |
>> http://www.securiteam.com/securitynews/5NP0C153PI.html |
| 33 |
> |
| 34 |
> kind regards! |
| 35 |
> |
| 36 |
sorry if I seemed rude. heh, text can come across like that sometimes. |
| 37 |
-- |
| 38 |
gentoo-performance@l.g.o mailing list |
Archives
