[rescue] You see a wild SPARCstation 20. What do?

[Kevin Bowling]

I'd be kind of surprised if two hme cards with an SM81 or late model
HS couldn't forward at close to full wire speed.  There are a number
of things to take into consideration.. network card design, driver
quality, bus speed, memory bandwidth, etc but still.. that's only
roughly 20MB/s even if it injects a bit more latency than a modern
system.

What hasn't been said, however, is that putting the SS20 into active
duty risks a neat machine.  If you're okay with losing a power supply
or whatever and have backups, go for it.  Otherwise,e keep it as a
hobby box to play with older Solaris versions and get a low power ARM
or x86 system to handle your routing.

Regards,
Kevin

On Sat, Jan 14, 2012 at 3:08 PM, Patrick Giagnocavo <xemacs5 at gmail.com>
wrote:
> My view is that the issue is memory latency.  Read/write speed to RAM
> matters a lot in such cases.
>
> I have seen this with Sparc and Perl processes also - the initial Perl
> startup is horrendous, then everything runs fine.
>
> Just sayin'
>
> Patrick
>
>> However, NAT and filtering add a fair amount of CPU-bound latency.  At one
>> time, my gateway was an Ultra 1e m170, and the combination of NAT +
>> filtering + PPPoE made performance absolutely awful.
> _______________________________________________
> rescue list - http://www.sunhelp.org/mailman/listinfo/rescue