On 1/18/07, Jeff Anderson-Lee <[EMAIL PROTECTED]> wrote:

Rick McNeal wrote:
> On Jan 18, 2007, at 12:00 PM, Mark A. Carlson wrote:
>
>> I would imagine the argument is that of software drivers
>> for the storage stack consume much less overhead by
>> cutting out the IP part - less CPU consumed, perhaps
>> better throughput. As far as cost, it leverages the NIC
>> commodity pricing curve without requiring TCP offload.
>
> The amount of CPU speed being consumed is really only valid for
> underpowered machines. Any modern desktop has more then enough
> horsepower to completely fill a 1GbE link with traffic at 4KB packet
> sizes.
I don't have OpenSolaris numbers at hand, but under Linux x86_64 on a
server-class motherboard that doesn't seem to be the case.

In a recent "echo"-style test, 4KiB UDP pegged one CPU of a dual 3.6GHz
Xeon EM64T but only obtained 90MB/s for UDP and 63MB/s for TCP/IP (with
no IPSEC).  It didn't saturate the network until sending 16KiB packets
for UDP and never did for TCP.  Perhaps with a TCP offload engine under
Solaris one might do better but... that's a lot of CPU power devoted
just to flinging the bits.


I just tested a quick rcp, and got about 110MB/s between two twin UltraSparc
machines, for the cost of about 2/3 of a 1.5GHz cpu. I know that pretty well
anything over 1GHz can saturate a 1GbE link - I've seen well
over 100MB/s serving NFS and doing backups and the machines are
hardly breaking into a sweat. That's entry level hardware with no
tweaking. Needless to say I have no idea how fast my Opteron boxes
could go as I simply don't have the network capacity to test it (over
120MB/s and they're hardly even getting warmed up).

--
-Peter Tribble
http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to