<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Colin Anderson wrote:
<blockquote
cite="mid03224271100A29458F0C144F1DB5304D0DBB@assprelay.landmarkhomes.net"
type="cite">
<pre wrap=""><!---->I'm quite aware that Opterons et al have gained signifigant market share in
the past couple years and AMD and their supporting chipsets have
dramatically improved in quality. I'm actually an AMD fan. However, when you
are spec'ing a system that a business will depend on (and your Asterisk
server is arguably the most important piece of kit in the rack) why would
you introduce an unknown variable in the equation just to save, say, 30% on
the price of the chip and motherboard? </pre>
</blockquote>
<snip><br>
<br>
The only comment I have in the same line as your comment is this.
Anyone who just goes out and buys a piece of hardware to run asterisk
on for thier business and does so, because it's a "tier 1" or reputable
manufacturer falls in my category of "gambler". If you haven't taken
the hardware you plan to build your business phone system on, and
installed the software it will be running and beat the ever living snot
out of it on a test bench, then you are asking for a suprise.<br>
<br>
This smells of the "You never get fired for buying IBM (replace with
Cisco, etc)" quote. Sure you can take the small gamble that a "tier 1"
platform will meet your needs, however, nothing beats a full battery of
tests including burn in, capacity and failure mode tests. If you don't
know how your system is going to fail and when it will fail, then how
are you going to monitor it?<br>
<br>
All in all, had Jenny's contractor done this, she would be in much
better shape now, however given the fact they didn't much care that the
card was sharing an IRQ with the network interface.....<br>
<br>
-Chris<br>
</body>
</html>