netdev
[Top] [All Lists]

RE: The ultimate TOE design

To: "'Nivedita Singhvi'" <niv@xxxxxxxxxx>
Subject: RE: The ultimate TOE design
From: "Leonid Grossman" <leonid.grossman@xxxxxxxx>
Date: Wed, 22 Sep 2004 21:46:47 -0700
Cc: "'Andi Kleen'" <ak@xxxxxxx>, "'David S. Miller'" <davem@xxxxxxxxxxxxx>, "'John Heffner'" <jheffner@xxxxxxx>, <netdev@xxxxxxxxxxx>, <linux-kernel@xxxxxxxxxxxxxxx>
In-reply-to: <4151DDFF.6000902@xxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
Thread-index: AcSg4Wls/1uX1Fa0Q4apuZMmesL9gQADYFIQ
> > 
> > It's a bit painful to setup, but in general with 9k jumbos 
> and TSO we 
> > were able to get close to pci-x 133 limit - both in LAN and 
> WAN tests.
> > Leonid
> 
> Cool, but a very specific environment, no? ;)

Define specific environment :-). We are running common tcp benchmarks like
nttcp or iperf or Chariot or filesystem applications on a very generic white
boxes, with generic OS/settings.

> 
> What concerns me about all this is that it seems so very 
> host-centric design. Wouldn't it be nice if we had a little 
> bit more network-centric worldview when designing network 
> infrastructure?
> 
> It isn't just a matter of how had we can push stuff out, it 
> also matters how much the network can take.
> Blasting tens of gigs into the ether seems all very exciting 
> sexy and cool, but suited for dedicated links or network 
> attached storage channels, not general-purpose networking on 
> the Internet or intra-nets.

This is somewhat different from IB or FC "miniature networks", 
some/most of 10GbE testing runs in existing datacenters or over 
existing long-haul links - see for example
http://sravot.home.cern.ch/sravot/Networking/10GbE/LSR_041504.htm

Cheers, Leonid

> 
> And if that is the case, we're talking about a much smaller 
> market (but perhaps a more profitable one ;))...
> 
> thanks,
> Nivedita
> 
> 
> 


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