LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: One-packet scheduling for kernel 2.6

To: Julian Anastasov <ja@xxxxxx>
Subject: Re: One-packet scheduling for kernel 2.6
Cc: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Cc: Wensong Zhang <wensong@xxxxxxxxxxxx>
From: Horms <horms@xxxxxxxxxxxx>
Date: Mon, 3 Oct 2005 19:23:41 +0900
On Mon, Oct 03, 2005 at 01:18:43PM +0300, Julian Anastasov wrote:
> 
>       Hello,
> 
> On Mon, 3 Oct 2005, Horms wrote:
> 
> > > - schedule packets according to Call-ID even if cp already exists
> > > - same CIP:CPORT can create many connections to different RS but to
> > > reduce their number they must be unique for RS->CIP:CPORT
> >
> > I would have thought you want to schedule packets based
> > on CIP:CPORT:Call-ID or perhaps just Call-ID
> >
> > Isn't the aim to make sure packets with the same Call-ID
> > end up at the same real-server?
> 
>       I assume requests will be forwarded just by Call-ID to
> allow single CIP:CPORT to use many RSs, while replies use the hashed
> conns to route replies back to CIP:CPORT. For NAT this ip_vs_nat_sip 
> module should be smarter, i'm not sure who will try to implement it, not 
> me :) I think, some IPs should be translated. For DR/TUN it looks easy,
> just ip_vs_sip.c scheduler.

Anything that involves NAT and SIP seems hard to me. The
implimentations I saw when playing with this about a year ago really
wanted end-to-end ness. Unfortunately I didn't have enough experience
with SIP, or with finding working implementations, to ever
satisfactorarily get SIP+LVS working at the time. Perhaps the wider
audience of this list can bring some skills to the table to make that
happen.  But to be honest, if we are just talking about load balancing
the control packets, is there really an application?

-- 
Horms

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