[89743] in North American Network Operators' Group

home help back first fref pref prev next nref lref last post

Re: OT: Xen

daemon@ATHENA.MIT.EDU (Eric Frazier)
Mon Apr 3 15:03:50 2006

Date: Mon, 03 Apr 2006 12:05:25 -0700
To: Michael.Dillon@btradianz.com, nanog list <nanog@merit.edu>
From: Eric Frazier <eric@dmcontact.com>
In-Reply-To: <OF9B5FBEB1.AFF599F2-ON80257145.00681756-80257145.00687C97@
 btradianz.com>
Errors-To: owner-nanog@merit.edu


At 12:01 PM 4/3/2006, Michael.Dillon@btradianz.com wrote:

> > Xen's bigges strength really is in the colocation business.  With
>VX-enabled
> > machines, it is capable of running instrumented OS's (Linux,
>Free/NetBSD) at
> > almost native speeds, and non-instrumented OS's (Windows, Solaris) with
>a
> > couple-% hit.  It's that flexibility that leads to colo as the market
>where
> > Xen shines.
>
>People seem to be thinking that Xen is only for sharing
>a colo machine with somebody else. But it could just as
>well be used for one organization to isolate each major
>application to a single virtual server, i.e. email server,
>general web server, wiki server, hot web app server,
>Asterisk server, etc. This way, when one of the applications
>justifies its own server, migration is somewhat simpler
>because it is not entangled with other applications.

Now that is what I have in mind. For me this is esp important where I have 
something nasty like a guy hosting a bunch of forums that are always not 
getting updated and getting defaced or worse. Until now I have had a dirty 
machine for stuff I know could lead to problems like that. But that brings 
up another question, how far isolated are different instances from each 
other really?



>-- Michael Dillon


home help back first fref pref prev next nref lref last post