[106469] in North American Network Operators' Group
Re: Great Suggestion for the DNS problem...?
daemon@ATHENA.MIT.EDU (Laird Popkin)
Tue Jul 29 12:25:51 2008
From: Laird Popkin <laird@pando.com>
To: Mikael Abrahamsson <swmike@swm.pp.se>
In-Reply-To: <alpine.DEB.1.10.0807291818440.7740@uplift.swm.pp.se>
Date: Tue, 29 Jul 2008 12:25:32 -0400
Cc: nanog@nanog.org
Errors-To: nanog-bounces@nanog.org
We mainly use UDP for tracker announces, and only use TCP when we have
to, and can confirm that the server spends far more time on the TCP
setup/teardown than on computing the tracker response.
- LP
On Jul 29, 2008, at 12:21 PM, Mikael Abrahamsson wrote:
> On Tue, 29 Jul 2008, Steven M. Bellovin wrote:
>
>> In this situation, UDP uses one query packet and one reply. TCP
>> uses 3
>> to set up the connection, a query, a reply, and three to tear down
>> the
>> connection. *Plus* the name server will have to keep state for
>> every client, plus TIMEWAIT state, etc. (Exercise left to TCP geek
>> readers: how few packets can you do this in? For example -- send the
>> query with the SYN+ACK, send client FIN with the query, send server
>> FIN
>> with the answer? Bonus points for not leaving the server's side in
>> TIMEWAIT. Exercise for implementers: how sane can your stack be if
>> you're going to support that?)
>
> The bittorrent tracker guys seem to run into problems at around 30kk
> tracker requests per second (TCP), and they say it's mostly setup/
> teardown (sy usage in vmstat), the tracker hash lookup doesn't take
> that much.
>
> They're trying to move to UDP, currently their workload is approx 5%
> UDP.
>
> I guess TCP DNS workload would be similar in characteristics.
>
> --
> Mikael Abrahamsson email: swmike@swm.pp.se
>
