djbdns
[Top] [All Lists]

Re: djbdns-1.05-epoll + speedup patch

To: djbdns Mailing List <dns@list.cr.yp.to>
Subject: Re: djbdns-1.05-epoll + speedup patch
From: Sami Farin <safari-dns@safari.iki.fi>
Date: Mon, 5 Mar 2007 21:26:01 +0200
Delivered-to: sp-com-lists@consult.net
Delivered-to: gmail-djbdns@securepoint.com
Delivered-to: sp.com.list@gmail.com
Delivered-to: mailing list dns@list.cr.yp.to
In-reply-to: <20070305164701.GA28354@odin.dempsky.org>
Mail-followup-to: djbdns Mailing List <dns@list.cr.yp.to>
Mailing-list: contact dns-help@list.cr.yp.to; run by ezmlm
References: <20070116171333.GA5674@m.safari.iki.fi> <20070222122129.GE3982@m.safari.iki.fi> <20070222162522.GG3982@m.safari.iki.fi> <20070304161657.qrytvzoi37inmj45@m.safari.iki.fi> <20070305142437.GA27256@odin.dempsky.org> <20070305144746.rrksiglhg3ljhvjp@m.safari.iki.fi> <20070305161316.GB27858@odin.dempsky.org> <20070305162742.GA15295@rfc822.net> <20070305164701.GA28354@odin.dempsky.org>
User-agent: Mutt/1.5.14 (2007-02-12)
On Mon, Mar 05, 2007 at 10:47:01 -0600, Matthew R. Dempsky wrote:
> On Mon, Mar 05, 2007 at 10:27:42AM -0600, Pete Ehlke wrote:
> > On Mon Mar 05, 2007 at 10:13:16 -0600, Matthew R. Dempsky wrote:
> > The important operational question is not the efficiency or correctness
> > of randomization in this area. The important question is the effect that
> > randomization and retry strategies have when faced with unreliable
> > networks and slow servers.
> > 
> > Compare dnscache with BIND 9 ( which has a more efficient retry and
> > selection mechanism ) in
> > http://dns.measurement-factory.com/writings/wessels-pam2004-paper.pdf
> 
> How is this relevant at all?  A quick review of the article indicates
> it discusses the pros and cons of randomly choosing nameservers to
> query rather than choosing them based on other heuristics.
> 
> The patch snippet in question does not change whether dnscache uses
> randomization---it merely (and seemingly unintentionally) changes the
> distribution of permutations from uniform to non-uniform.

Why do you think it was unintentional?
Why do you think distribution of permutations matters in this case?

-- 

Attachment: djbdns-1.05-epoll-20070305T031624Z-20070305T185643Z.diff
Description: Text document

Attachment: pgp32fwFARahM.pgp
Description: PGP signature

<Prev in Thread] Current Thread [Next in Thread>