<font class="Apple-style-span" face="arial, helvetica, sans-serif"><br></font><div class="gmail_quote">On 4 April 2011 12:31, Soren Hansen <span dir="ltr"><<a href="mailto:soren@linux2go.dk">soren@linux2go.dk</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div id=":1ev">2011/4/2 Serge van Ginderachter <<a href="mailto:serge@vanginderachter.be">serge@vanginderachter.be</a>>:<br>
<div class="im">> On 2 April 2011 16:58, Clint Byrum <<a href="mailto:clint@ubuntu.com">clint@ubuntu.com</a>> wrote:<br>
</div><div class="im">>> Serge, would you mind elaborating on that? I'm looking for facts.<br>
><br>
> I tested several virtualisation technologies last year.<br>
<br>
</div>I'm terribly sorry, but this information is practically useless. There<br></div></blockquote><div><br></div><meta http-equiv="content-type" content="text/html; charset=utf-8">Yes, I very well realise that. <meta http-equiv="content-type" content="text/html; charset=utf-8">I only gave some general concusions to give an idea of what we were looking at.<div>
I was only trying to slightly elaborate on the matter as an answer to an earlier question on the list.</div><div><br></div><div>Unfortunately, the reports of those tests are private and not published, and I'm not allowed to do that (don't ask), so I can't fully disclose them.</div>
<div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div id=":1ev">
are no version numbers, no information about configuration, about</div></blockquote><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div id=":1ev">
backing stores, disk image formats, cache settings, and very little<br>
about hardware, etc. I can't e.g. tell if your factor 8 drop i<br>
performance on Ubuntu for small writes is due to the virtual disk<br>
being backed by a qcow2 on ext4, for instance, and as such, I can't<br>
use the data (and much less the conclusions) for anything.</div></blockquote></div><br><div>Yes, lots of things could be optimised, that's for sure. But the main aim of the tests were primarily about comparing Xen and KVM, and as such, similar setups (LVM backed disks, Virtio/HVM hardware, using the same startup scripts on all platforms, ... ) and pretty much most default settings were used. The VM images were all identical, Debian Lenny with ext3.</div>
<div><br></div><div>So, while different settings might not be fully optimized in those tests - at the time we were pretty new with this stuff - we did made several tests which could compare different platforms. And our conclusion to that was that KVM in general was less performant than Xen.</div>
<div><br></div><div>That is the only point I wanted to make. Obviously, YMMV. </div><div><br></div><div>Also note that these conclusions don't stop me from still using Ubuntu+KVM for lots of setups, but more because of ease of use than performance.</div>
<div><br>-- <br><div><font><font><div><font><div><font><font><div><font><div><div style="white-space:pre-wrap"><font face="georgia, serif"><span style="font-family:arial;white-space:normal"><div style="white-space:pre-wrap">
<font color="#666666" face="'arial narrow', sans-serif"><span style="font-size:x-small"><span style="font-family:arial, helvetica, sans-serif"><div style="white-space:pre-wrap"><span style="white-space:normal"><font face="'arial narrow', sans-serif">Met vriendelijke groet,</font></span></div>
<div><font face="'arial narrow', sans-serif">Serge van Ginderachter<br></font></div></span></span></font></div></span></font></div><div><div style="white-space:normal;font-family:arial, helvetica, sans-serif"><font face="arial,helvetica,sans-serif"></font></div>
</div></div></font></div></font></font></div></font></div></font></font></div><br>
</div>