[rescue] Perverse Question
mcguire at neurotica.com
Mon Jun 16 15:33:07 CDT 2003
On Monday, June 16, 2003, at 05:52 AM, Al Potter wrote:
>> It's time to let go of that. I currently run no Sun4c or Sun4m
>> machines in production. I merely stated, repeatedly and correctly,
>> that an SS2 (as a web server) is easily able to saturate any
>> connectivity that anyone here is likely to have.
> It's finally hit me....
> Yes, a SS2 as a web server can saturate any pipe anyone of us is
> likely to
> afford, but....
> - I suspect you think of a web server like this and mean "serving
> pages", not running some dynamic backend or server-side include lashup
> seems to be the rage today. You might still achieve saturation, but
> not, I
> would suspect, with acceptable latency.
You make an excellent point as always, but...
No, I was talking about static pages. Though they're not as
"glamorous" these days, most of the pages on the web are static.
A dynamically-generated page could be anything from a server-side
include which sticks a timestamp in a page or something that passes
protein-folding simulation parameters to a cluster of Crays and
displays the results. Where do we draw the line? ;)
> - remember, a PC with an 8bit bic could saturate ethernet. This did
> meat it was useful while doing so.
I've *never* seen a PC ISA NIC that could saturate a 10mbit ethernet.
In fact, the first machine I did see which could was an SGI Personal
Besides, though...we're not talking about just the network interface
being able to push the bits down the wire (well, at least that's not
what I was talking about..) I was talking about
systems-level...Network interface, processor, memory, disk subsystem,
OS, server software...the whole deal. Which, for static pages and
Apache running under something other than Solaris, an SS2 can do all
day and all night without even showing any load.
Dave McGuire "I've grown hair again, just
St. Petersburg, FL for the occasion." -Doc Shipley
More information about the rescue