Re: FTP benchmark proposal

Paul M. Hirsch (pauldoom@telebot.com)
Mon, 28 Jun 1999 13:11:50 -0500 (CDT)


Everyone -- Great ideas in this thread. Linux needs at least a few high
perf benchmark centers that are permanent fixtures, instead of one time
only shoot-out type set ups. In order to slay NT, (as we shall ;), we need
some accessable high performance test environments for kernel code.

In reference to your idea about asking ZD to help, I think we should be
wary. ZD is a news organization. Whatever we do on thier dollar would be
for news, not necissarily the good of Linux, no matter what thier
intentions (which actually seem to be friendly).

> a word on hardware: we _definitely_ want to use a gigabit rather than a
> pile of 100mb cards. gigabit cards are cheapish when compared to the
> drives and memory and such in the machine, and aggregate lots of traffic
> infinitely better than a pile of 100mb cards. (the exception being the

We should _definitely_ have both scenarios available. M$ seems to have
found a weakness in our network code with multiple NICs. We shouldn't
just turn to Gbit as the answer. We (actually, others who know a hell of
a lot more than I) need to fix the multi NIC problem or it will show up
again when people start putting multi Gbit cards in thier 21364 Alphas.
The truth is that lots of folks use multi 100Mb instead of Gb. If Linux
gets beat by NT in that situation, that is bad and should be fixed.

That said, I look forward to seeing your FTPD pulling 20MB/sec+
sustained transfer. :)

-Paul

----------------------------------------------------------------------
Paul M. Hirsch Network Engineer EXi, Corp.
smtp[0]-->phirsch@exicorp.com smtp[1]-->pauldoom@telebot.com
vox.mobile-->612.414.5447 pager-->612.909.8762
PGP fingerprint = CF43 3528 BA9D 9681 C242 06C3 6235 8573 71EA 5828
----------------------------------------------------------------------

-
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/