Re: Gigabit Linux Server Bottlenecks

From: Jes Sorensen (Jes.Sorensen@cern.ch)
Date: Tue Feb 08 2000 - 07:16:20 EST


>>>>> "Anton" == Anton Ivanov <aivanov@eu.level3.net> writes:

>> This is correct, but it doesn't change the fact that you still
>> need to go through the number of IP and TCP headers, having larger
>> packets reduces the amount of work the host CPU has to do
>> significantly. If

Anton> Most cards have rudimentary IP stack functions on them as
Anton> well. At least PacketEngines and methings Intel should be able
Anton> to do the checksumming and a few other rudimentary ops.

Anton> These are not implemented in the Linux driver (maybe for better
Anton> ;-).

Alteon's AceNIC does TCP checksumming, I utilize it in my driver for
receive. There is no point in using it for transmit since the Linux
TCP stack currently doesn't support this.

Anton> There is still a few things that it makes sense for: 802.1Q
Anton> being the most noteable example. Overcoming the machine port
Anton> number limitations and having more than 10 Ethernet interfaces
Anton> does make sense. Any chance for it making mainstream?

I still haven't figure out what I need 802.1Q for, but maybe thats
just me. Guess I ought to go read the standard.

Jes

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Feb 15 2000 - 21:00:12 EST