LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: [lvs-users] Re: IPVS Benchmarking

To: Joseph Mack <mack@xxxxxxxxxxx>, Julian Anastasov <uli@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: [lvs-users] Re: IPVS Benchmarking
Cc: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
From: Wayne <wayne@xxxxxxxxxxxxxxx>
Date: Mon, 24 Jan 2000 08:53:39 -0800
At 07:15 AM 1/23/00 -0500, Joseph Mack wrote:
>On Sun, 23 Jan 2000, Julian Anastasov wrote:
>
> >       The wrong in your setup is that in VS/NAT mode the Director is
> > used with 1 NIC. Assume that the requests from the client and the answers
> > from the real servers use the same link twice:
>
>It's a good idea to test VS-NAT with 2 NICs to optimise the hardware for
>each type of LVS. For my first test I used one NIC so that I could compare
>VS-NAT and VS-DR on the same hardware.
>
> > What about testing input requests with size = 10-20% the size
> > of the answers (httpd traffic), what is the difference between VS/NAT and
> > VS/DR even when the Director is with one NIC? 
>
>yes all of these tests need to be done too.
>
> >       What means throughput 120 Mbps for the real servers in the table?
> > Is the limit 100Mbps for the NICs? Input+Output in Full-Duplex ? I don't
> > understand something in this table. You report that the throughput from
> > the client to the real server (directly?) is 50Mbps. If the size of the
> > request = size of the answer, the max throughput can be 50Mbps (reported
> > in the real server, half-duplex).
>
>Yes this is a puzzle and I don't understand it myself. The connection
>client-director is 70Mbps by netpipe for a single netpipe run. I assume
>that means 70Mbps in each direction and that the total bit rate is
>140Mbps. I assume that the cards are all limited to 100Mbps by all having
>to use the same clock speed so that they can talk to each other. I can
>imagine that some cards might have slower average throughput because of
>buffer filling etc, but it's hard to imagine a situation where the speed
>would go above 100Mbps. I don't understand then why the client then can
>register a total of 120Mbps on netpipe (6 windows each registering
>20Mbps). I should do the same test connecting the client directly to the
>director running 6 netpipe sessions.
>
>Do you know if a 100Mpbs NIC is supposed to be able to receive at 100Mbps
>at the same time as it is sending 100Mpbs? Does an 8-port switch have 4
>pairs of connections in both directions at 100Mbps (total 800Mpbs) or only
>4 connections with each connection being in 1 direction (total 400Mbps)?

Intel EtherExpress 100+ cards work reliably at full-duplex mode, which
is 100MBS up and 100MBS down at the same time.  Intel's cards also
support merge bandwidth, that is two cards together it could achieve
200MBS up and 200MBS down.  Have not tested the Linux version
driver to see if that was working or not with Linux.


>Joe
>--
>Joseph Mack mack@xxxxxxxxxxx
>
>
>----------------------------------------------------------------------
>LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
>To unsubscribe, e-mail: lvs-users-unsubscribe@xxxxxxxxxxxxxxxxxxxxxx
>For additional commands, e-mail: lvs-users-help@xxxxxxxxxxxxxxxxxxxxxx


----------------------------------------------------------------------
LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
To unsubscribe, e-mail: lvs-users-unsubscribe@xxxxxxxxxxxxxxxxxxxxxx
For additional commands, e-mail: lvs-users-help@xxxxxxxxxxxxxxxxxxxxxx

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