LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: Is active-active LVS really capable of doubling theoretical bandwid

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: Is active-active LVS really capable of doubling theoretical bandwidth?
From: Joseph Mack NA3T <jmack@xxxxxxxx>
Date: Fri, 3 Nov 2006 05:07:17 -0800 (PST)
On Thu, 2 Nov 2006, Michael Spiegle wrote:

I did some reading on active-active and while I think its a great idea,
I didn't quite understand how it was capable of doubling the throughput
as mentioned in the PDF document.

Horms would be the final word on the matter, but he seems to be busy (hasn't popped up on the mailing list for a few weeks). I haven't read the Saru docs for a couple of years and you seem to have read them thoroughly and I assume you've got it right.

n active directors can handle n times the bandwidth, but you have to send it to them. You'll need 10GigE (or something like that) in front of them.

I didn't know that the directors had iptables rules infront of them to divide the work load. One of the requirements is to load balance and handle failover, so there must be at least two directors capable of handling any particular stream of traffic.

Joe

--
Joseph Mack NA3T EME(B,D), FM05lw North Carolina
jmack (at) wm7d (dot) net - azimuthal equidistant map
generator at http://www.wm7d.net/azproj.shtml
Homepage http://www.austintek.com/ It's GNU/Linux!

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