[178179] in North American Network Operators' Group
Re: OT - Small DNS "appliances" for remote offices.
daemon@ATHENA.MIT.EDU (Rob Seastrom)
Thu Feb 19 06:18:47 2015
X-Original-To: nanog@nanog.org
To: Bryan Seitz <seitz@bsd-unix.net>
From: Rob Seastrom <rs@seastrom.com>
Date: Thu, 19 Feb 2015 06:18:43 -0500
In-Reply-To: <20150219035155.GA13222@bsd-unix.net> (Bryan Seitz's message of
"Wed, 18 Feb 2015 22:51:55 -0500")
Cc: "nanog@nanog.org >> 'NANOG list'" <nanog@nanog.org>
Errors-To: nanog-bounces@nanog.org
Bryan Seitz <seitz@bsd-unix.net> writes:
> odroid-c1 + eMMC module + RTC battery + case + power adapter.
> Should run you about $75 *AND* wouldn't be bad for running NTP as
> well.
I haven't looked into the details of the clock, so "wouldn't be bad"
is probably true, "notably good", well, that would be a task for
someone with experience doing clock benchmarking and who can describe
MAVAR without looking it up.
> The gig-e port on the C1 has been observed to push 405Mbps TX and
> 940Mbps+ RX via iperf.
The 405 Mbps for TX. I've seen around 30 Mbyte/sec on single stream
TCP RX. Got 99.5 Mbyte/sec from a Mac Mini in the same subnet so
that's not a limit of the host on the other end of the benchmark.
I call shenanigans on the 940 Mbps iperf number though. The HSIC bus
is only 480 Mbit/sec. Two pints of beer in a one pint glass would be
some trick.
-r
