LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: Keepalived + feedbackd

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: Keepalived + feedbackd
From: Alexandre Cassen <Alexandre.Cassen@xxxxxxx>
Date: Fri, 07 Oct 2005 00:36:49 +0200
Hello,

At 15:56 06/10/2005, Jacob Coby wrote:
Jeremy Kerr wrote:
There were discussions along these lines a while ago - we decided that it would be the right thing to do, but was put "on the back-burner", as there doesn't seem to be a whole lot of interest.
What do you guys think? Is there renewed interest in active load monitoring?

I'd be willing to help. I started on it last month, but didn't continue because I never heard back from Alexandre.


Sorry for delay, currently too busy. I spoke with cisco guy about their feedback engine in their SLB card, and they are pushing the SASP protocol ahead for large datacenter :

        http://www.ietf.org/internet-drafts/draft-bivens-sasp-02.txt

Why not using SASP instead of NECP ?

OTOH, I really think if we want to produce scalable and fault tolerant solutions for a feedback channel between director(s) and realserver, then we MUST use multicast datagram IMHO...

Best regards,
Alexandre


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