To: Horms <horms@verge.net.au>
From: Theodore Tso <tytso@mit.edu>
Cc: evals@lists.merlins.org
Date: Thu, 8 Feb 2007 21:32:38 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209005801.GB2847@verge.net.au>
Message-Id: <20070209023238.GE13207@thunk.org>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>

On Fri, Feb 09, 2007 at 09:58:02AM +0900, Horms wrote:
> Parts of XenSource's code may be opensource, but its certainly
> not a community based project. By which I mean, most people
> work on the code because they are told to by their boss, not
> because they want to or have any interest in working on it
> in their spare time. Its feature and release driven more than
> quality and engineering driven.
> 
> I'm not saying that there is anything wrong with this,
> but the energy is certainly different.

.... and they have been completely lame about trying to get Xen merged
into Linux mainline, with the result that Xen is about to be rendered
completely irrelevant by KVM in 6-9 months.

I pity the poor distributions that got suckered by their marketing
folks to include Xen into their distro's in violation of their general
policy of waiting until said technology gets merged into mainline.

As a result I have to wonder about how viable XenSource will be as a
company a year or so from now.

							- Ted

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: Theodore Tso <tytso@mit.edu>
From: Horms <horms@verge.net.au>
Cc: evals@lists.merlins.org
Date: Fri, 9 Feb 2007 12:27:32 +0900
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: mutt-ng/devel-r804 (Debian)
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209023238.GE13207@thunk.org>
Message-Id: <20070209032731.GA5709@verge.net.au>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>

On Thu, Feb 08, 2007 at 09:32:38PM -0500, Theodore Tso wrote:
> On Fri, Feb 09, 2007 at 09:58:02AM +0900, Horms wrote:
> > Parts of XenSource's code may be opensource, but its certainly
> > not a community based project. By which I mean, most people
> > work on the code because they are told to by their boss, not
> > because they want to or have any interest in working on it
> > in their spare time. Its feature and release driven more than
> > quality and engineering driven.
> > 
> > I'm not saying that there is anything wrong with this,
> > but the energy is certainly different.
> 
> .... and they have been completely lame about trying to get Xen merged
> into Linux mainline, with the result that Xen is about to be rendered
> completely irrelevant by KVM in 6-9 months.
> 
> I pity the poor distributions that got suckered by their marketing
> folks to include Xen into their distro's in violation of their general
> policy of waiting until said technology gets merged into mainline.
> 
> As a result I have to wonder about how viable XenSource will be as a
> company a year or so from now.

Agreed, 100%

-- 
Horms
  H: http://www.vergenet.net/~horms/
  W: http://www.valinux.co.jp/en/


_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: Horms <horms@verge.net.au>
From: gregory pryzby <greg@pryzby.org>
Cc: evals@lists.merlins.org
Date: Thu, 8 Feb 2007 22:45:12 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209032731.GA5709@verge.net.au>
Message-Id: <20070209034512.GI28750@pryzby.org>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>

On Fri, Feb 09, 2007 at 12:27:32PM +0900, Horms wrote:
> On Thu, Feb 08, 2007 at 09:32:38PM -0500, Theodore Tso wrote:
> > On Fri, Feb 09, 2007 at 09:58:02AM +0900, Horms wrote:
> > > Parts of XenSource's code may be opensource, but its certainly
> > > not a community based project. By which I mean, most people
> > > work on the code because they are told to by their boss, not
> > > because they want to or have any interest in working on it
> > > in their spare time. Its feature and release driven more than
> > > quality and engineering driven.
> > > 
> > > I'm not saying that there is anything wrong with this,
> > > but the energy is certainly different.
> > 
> > .... and they have been completely lame about trying to get Xen merged
> > into Linux mainline, with the result that Xen is about to be rendered
> > completely irrelevant by KVM in 6-9 months.

is KVM all that? speed, performance, etc?

The management tools are what is needed for ANY virtualization
solution.

> > I pity the poor distributions that got suckered by their marketing
> > folks to include Xen into their distro's in violation of their general
> > policy of waiting until said technology gets merged into mainline.
> > 
> > As a result I have to wonder about how viable XenSource will be as a
> > company a year or so from now.
> 
> Agreed, 100%

-- 
greg pryzby                              greg at pryzby dot org
fingerprint: 8A1A DB90 869F 5DD1 D6E9 EEB6 C156 6B04 849F A86F
_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals

To: gregory pryzby <greg@pryzby.org>
From: Horms <horms@verge.net.au>
Cc: evals@lists.merlins.org
Date: Fri, 9 Feb 2007 14:08:08 +0900
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: mutt-ng/devel-r804 (Debian)
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209034512.GI28750@pryzby.org>
Message-Id: <20070209050806.GA6973@verge.net.au>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>

On Thu, Feb 08, 2007 at 10:45:12PM -0500, gregory pryzby wrote:
> On Fri, Feb 09, 2007 at 12:27:32PM +0900, Horms wrote:
> > On Thu, Feb 08, 2007 at 09:32:38PM -0500, Theodore Tso wrote:
> > > On Fri, Feb 09, 2007 at 09:58:02AM +0900, Horms wrote:
> > > > Parts of XenSource's code may be opensource, but its certainly
> > > > not a community based project. By which I mean, most people
> > > > work on the code because they are told to by their boss, not
> > > > because they want to or have any interest in working on it
> > > > in their spare time. Its feature and release driven more than
> > > > quality and engineering driven.
> > > > 
> > > > I'm not saying that there is anything wrong with this,
> > > > but the energy is certainly different.
> > > 
> > > .... and they have been completely lame about trying to get Xen merged
> > > into Linux mainline, with the result that Xen is about to be rendered
> > > completely irrelevant by KVM in 6-9 months.
> 
> is KVM all that? speed, performance, etc?

Think about this for a moment:

If you want to use virtualisation for server consolidation,
then presumably the servers you are consolodating are reasonably idle.
It seems that if that is the case, performance isn't the primary concern
at hand.

> The management tools are what is needed for ANY virtualization
> solution.

Yes. But you need to build on a solid foundation. And Xen smells
like the house built on the dunes that is wiped out in a storm.



-- 
Horms
  H: http://www.vergenet.net/~horms/
  W: http://www.valinux.co.jp/en/


_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: Horms <horms@verge.net.au>
From: gregory pryzby <greg@pryzby.org>
Cc: evals@lists.merlins.org
Date: Fri, 9 Feb 2007 06:43:45 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209050806.GA6973@verge.net.au>
Message-Id: <20070209114345.GA29824@pryzby.org>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209050806.GA6973@verge.net.au>

On Fri, Feb 09, 2007 at 02:08:08PM +0900, Horms wrote:
> On Thu, Feb 08, 2007 at 10:45:12PM -0500, gregory pryzby wrote:
> > On Fri, Feb 09, 2007 at 12:27:32PM +0900, Horms wrote:
> > > On Thu, Feb 08, 2007 at 09:32:38PM -0500, Theodore Tso wrote:
> > > > On Fri, Feb 09, 2007 at 09:58:02AM +0900, Horms wrote:
> > > > > Parts of XenSource's code may be opensource, but its certainly
> > > > > not a community based project. By which I mean, most people
> > > > > work on the code because they are told to by their boss, not
> > > > > because they want to or have any interest in working on it
> > > > > in their spare time. Its feature and release driven more than
> > > > > quality and engineering driven.
> > > > > 
> > > > > I'm not saying that there is anything wrong with this,
> > > > > but the energy is certainly different.
> > > > 
> > > > .... and they have been completely lame about trying to get Xen merged
> > > > into Linux mainline, with the result that Xen is about to be rendered
> > > > completely irrelevant by KVM in 6-9 months.
> > 
> > is KVM all that? speed, performance, etc?
> 
> Think about this for a moment:
> 
> If you want to use virtualisation for server consolidation,
> then presumably the servers you are consolodating are reasonably idle.
> It seems that if that is the case, performance isn't the primary concern
> at hand.

agreed they are not being used to their best. However, if I can
consolidate 10 to 1 vs 2 to 1, I am going to look at 10 to 1. Also
running a mixed environment is something customers want. If the
customer can't mix the OSs on the machine, they are not interested.

> > The management tools are what is needed for ANY virtualization
> > solution.
> 
> Yes. But you need to build on a solid foundation. And Xen smells
> like the house built on the dunes that is wiped out in a storm.

I am not smart enough to know the answer to that. i have asked why
upstream isn't happening, but I am asking the wrong people (people who
only guess and don't know).

What I have read (so it is worth most stock option grants) is the kvm
is an interesting start, but right now, it isn't ready to show or even
use in a beta. I have no idea if that is true.

And until there is a management tool, it really doesn't matter for
most companies of size. Because a vitualization is instances plus 1
when it comes to managing them. So more management.
 

-- 
greg pryzby                              greg at pryzby dot org
fingerprint: 8A1A DB90 869F 5DD1 D6E9 EEB6 C156 6B04 849F A86F
_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals

To: gregory pryzby <greg@pryzby.org>
From: Wichert Akkerman <wichert@wiggy.net>
Cc: evals@lists.merlins.org
Date: Fri, 9 Feb 2007 13:02:21 +0100
Subject: [Evals] SPAM: 8.9: Re:  Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.9i
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209114345.GA29824@pryzby.org>
Message-Id: <20070209120221.GS23685@wiggy.net>
References: <1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209050806.GA6973@verge.net.au>
	<20070209114345.GA29824@pryzby.org>

Previously gregory pryzby wrote:
> On Fri, Feb 09, 2007 at 02:08:08PM +0900, Horms wrote:
> > Think about this for a moment:
> > 
> > If you want to use virtualisation for server consolidation,
> > then presumably the servers you are consolodating are reasonably idle.
> > It seems that if that is the case, performance isn't the primary concern
> > at hand.
> 
> agreed they are not being used to their best. However, if I can
> consolidate 10 to 1 vs 2 to 1, I am going to look at 10 to 1. Also
> running a mixed environment is something customers want. If the
> customer can't mix the OSs on the machine, they are not interested.

Also if you are in the hosting business you do not always want to sell
top performance. You want to sell cheap and offer top performance as
an extra service at extra cost.

Wichert.

-- 
Wichert Akkerman <wichert@wiggy.net>    It is simple to make things.
http://www.wiggy.net/                   It is hard to make things simple.

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: gregory pryzby <greg@pryzby.org>, Horms <horms@verge.net.au>,
	evals@lists.merlins.org
From: Wichert Akkerman <wichert@wiggy.net>
Date: Fri, 9 Feb 2007 13:03:50 +0100
Subject: [Evals] SPAM: 8.9: Re: SPAM: 8.9: Re: Zimbra: Am in San Mateo this
	weekend
User-Agent: Mutt/1.5.9i
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209120221.GS23685@wiggy.net>
Message-Id: <20070209120350.GT23685@wiggy.net>
References: <20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209050806.GA6973@verge.net.au>
	<20070209114345.GA29824@pryzby.org>
	<20070209120221.GS23685@wiggy.net>

What's up with the 'SPAM: 8.9' in the subject?

Wichert.

-- 
Wichert Akkerman <wichert@wiggy.net>    It is simple to make things.
http://www.wiggy.net/                   It is hard to make things simple.

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: gregory pryzby <greg@pryzby.org>
From: Theodore Tso <tytso@mit.edu>
Cc: evals@lists.merlins.org
Date: Fri, 9 Feb 2007 08:22:55 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209034512.GI28750@pryzby.org>
Message-Id: <20070209132255.GA9902@thunk.org>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>

On Thu, Feb 08, 2007 at 10:45:12PM -0500, gregory pryzby wrote:
> 
> is KVM all that? speed, performance, etc?

It will be; there already is code being worked on (not yet merged)
that will use paravirt_ops to speed up common operations, as well as
direct emulation drivers that a guest OS can use so that it doesn't
have to use the qemu hardware emulation layer.  (That will still be
provided if you want to run Legacy OS's like Windows, of course.)

So fairly soon, KVM will have all of the paravirtualization speed
advantages of Xen, in a codebase which is clean enough that kernel
developers aren't barfing all over it.  If XenSource had been willing
to suck it up and do the work necessary to merge it into mainline,
then maybe the VT technology could have been seen as a Xen
accelerator, but instead, it's going to be the other way around; KVM
will be able to run unmodified legacy OS's just fine (which Xen can
not), as well as running Linux guest OS's very quickly with
paravirtualization enabled.

> The management tools are what is needed for ANY virtualization
> solution.

... and truth be told, the management tools aren't really going to be
that hard to provide.  I predict that OSS tools will become more and
more real over the next year or two, now that we have an open,
mainline virtualization tool.  It may use system administrator
friendly command-line and scripting techniques instead of marketing
manager-friendly GUI's, at least at first, but most of the people who
are addicted to pretty GUI's are probably using VMware, anyway.  

So that's why I think XenSource will likely start losing out within
1-2 years.  The only hope they have is that the distribution's were
suckered into shipping Xen, and that KVM requires newer hardware.  But
over time, more and more hardware will have Vanderpool or VT support
in it, and XenSource is going to become more and more hosed....

						- Ted

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: Theodore Tso <tytso@mit.edu>
From: gregory pryzby <greg@pryzby.org>
Cc: evals@lists.merlins.org
Date: Fri, 9 Feb 2007 08:29:50 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209132255.GA9902@thunk.org>
Message-Id: <20070209132950.GA30104@pryzby.org>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209132255.GA9902@thunk.org>

So where can I read and learn more about KVM (besides reading the
code!)?

On Fri, Feb 09, 2007 at 08:22:55AM -0500, Theodore Tso wrote:
> On Thu, Feb 08, 2007 at 10:45:12PM -0500, gregory pryzby wrote:
> > 
> > is KVM all that? speed, performance, etc?

SNIP lots of information that is worth reading 

-- 
greg pryzby                              greg at pryzby dot org
fingerprint: 8A1A DB90 869F 5DD1 D6E9 EEB6 C156 6B04 849F A86F
_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals

To: evals@lists.merlins.org
From: James Antill <james-evals@and.org>
Date: Fri, 09 Feb 2007 10:19:04 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Gnus/5.1007 (Gnus v5.10.7) XEmacs/21.5-b27 (linux)
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209132255.GA9902@thunk.org> (Theodore Tso's message of "Fri,
	9 Feb 2007 08:22:55 -0500")
Message-Id: <m3tzxvcqaf.fsf@code.and.org>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org> <20070209132255.GA9902@thunk.org>

Theodore Tso <tytso@mit.edu> writes:

> On Thu, Feb 08, 2007 at 10:45:12PM -0500, gregory pryzby wrote:
>> The management tools are what is needed for ANY virtualization
>> solution.
>
> ... and truth be told, the management tools aren't really going to be
> that hard to provide.  I predict that OSS tools will become more and
> more real over the next year or two, now that we have an open,
> mainline virtualization tool.  It may use system administrator
> friendly command-line and scripting techniques instead of marketing
> manager-friendly GUI's, at least at first, but most of the people who
> are addicted to pretty GUI's are probably using VMware, anyway.  

 It's also worth noting that all of the tools for Fedora are written
using libvirt, specifically so it'd be easy for them to dump
Xe^H^H^H^H^H^H^H take advantage of any new upstream virtualization
technologies.

 I don't speak for anyone else, I'm not in the virt. engineering
groups at any company etc. etc.

-- 
James Antill -- james@and.org
http://www.and.org/and-httpd/ -- $2,000 security guarantee
http://www.and.org/vstr/

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: gregory pryzby <greg@pryzby.org>
From: Theodore Tso <tytso@mit.edu>
Cc: evals@lists.merlins.org
Date: Sat, 10 Feb 2007 11:36:49 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209132950.GA30104@pryzby.org>
Message-Id: <20070210163649.GD25607@thunk.org>
References: <1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209132255.GA9902@thunk.org>
	<20070209132950.GA30104@pryzby.org>

On Fri, Feb 09, 2007 at 08:29:50AM -0500, gregory pryzby wrote:
> So where can I read and learn more about KVM (besides reading the
> code!)?

Probably the best place to start is the KVM wiki:

http://kvm.qumranet.com/kvmwiki

						- Ted

P.S. This may be another sign of XenSource's doom; the KVM wiki is
being hosted by Qumranet, a startup currently in stealth mode.  But if
you check out http://www.qumranet.com, you'll see that that the CTO of
Qumranet is Moshe Bar, prior founder and CTO of XenSource.  If I
worked at XenSource, I would certainly be spiffing up my resume...


_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: "Theodore Tso" <tytso@mit.edu>
From: "Chris DiBona" <cdibona@gmail.com>
Cc: evals@lists.merlins.org
Date: Sat, 10 Feb 2007 10:30:54 -0800
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070210163649.GD25607@thunk.org>
Message-Id: <7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>
References: <1170950089.3916.9.camel@localhost.localdomain>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org> <20070209132255.GA9902@thunk.org>
	<20070209132950.GA30104@pryzby.org> <20070210163649.GD25607@thunk.org>

It would seem somewhat unethical to compete with your former company in
which you would have a founder level of stock. Maybe he is a fan of
virtualization? It's no secret that Xen went through a , ahem, reimagining
last year.

Chris

On 2/10/07, Theodore Tso <tytso@mit.edu> wrote:
>
> On Fri, Feb 09, 2007 at 08:29:50AM -0500, gregory pryzby wrote:
> > So where can I read and learn more about KVM (besides reading the
> > code!)?
>
> Probably the best place to start is the KVM wiki:
>
> http://kvm.qumranet.com/kvmwiki
>
>                                                 - Ted
>
> P.S. This may be another sign of XenSource's doom; the KVM wiki is
> being hosted by Qumranet, a startup currently in stealth mode.  But if
> you check out http://www.qumranet.com, you'll see that that the CTO of
> Qumranet is Moshe Bar, prior founder and CTO of XenSource.  If I
> worked at XenSource, I would certainly be spiffing up my resume...
>
>
> _______________________________________________
> Evals mailing list
> Evals@lists.merlins.org
> http://lists.merlins.org/lists/listinfo/evals
>



-- 
Open Source Programs Manager, Google Inc.
Google's Open Source program can be found at http://code.google.com
Personal Weblog: http://dibona.com
_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals

To: Ex-VA Linux Systems <evals@lists.merlins.org>
From: gregory pryzby <greg@pryzby.org>
Date: Sat, 10 Feb 2007 13:44:03 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>
Message-Id: <20070210184403.GB2341@pryzby.org>
References: <1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209132255.GA9902@thunk.org>
	<20070209132950.GA30104@pryzby.org>
	<20070210163649.GD25607@thunk.org>
	<7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>

On Sat, Feb 10, 2007 at 10:30:54AM -0800, Chris DiBona wrote:
> It would seem somewhat unethical to compete with your former company in which
> you would have a founder level of stock. Maybe he is a fan of virtualization?
> It's no secret that Xen went through a , ahem, reimagining last year.

I have no idea what he does or doesn't have. 

I do know a few people at XS who are 'not a fan' of Moshe. No idea
what happened. 
 
I think that Xen is not in mainstream will hurt them, I also think if
they got their act together and got accepted into the kernel, there
would be a chance for them.

Look at how SELinux is still around
 
> On 2/10/07, Theodore Tso <tytso@mit.edu> wrote:
> 
>     On Fri, Feb 09, 2007 at 08:29:50AM -0500, gregory pryzby wrote:
>     > So where can I read and learn more about KVM (besides reading the
>     > code!)?
> 
>     Probably the best place to start is the KVM wiki:
> 
>     http://kvm.qumranet.com/kvmwiki
> 
>                                                     - Ted
> 
>     P.S. This may be another sign of XenSource's doom; the KVM wiki is
>     being hosted by Qumranet, a startup currently in stealth mode.  But if
>     you check out http://www.qumranet.com, you'll see that that the CTO of
>     Qumranet is Moshe Bar, prior founder and CTO of XenSource.  If I
>     worked at XenSource, I would certainly be spiffing up my resume...

-- 
greg pryzby                              greg at pryzby dot org
fingerprint: 8A1A DB90 869F 5DD1 D6E9 EEB6 C156 6B04 849F A86F
_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals

To: evals@lists.merlins.org
From: Michael Jennings <mej+evals@kainx.org>
Date: Sun, 11 Feb 2007 15:28:16 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.11
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070210184403.GB2341@pryzby.org>
Message-Id: <20070211202816.GD7919@kainx.org>
References: <20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209132255.GA9902@thunk.org>
	<20070209132950.GA30104@pryzby.org>
	<20070210163649.GD25607@thunk.org>
	<7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>
	<20070210184403.GB2341@pryzby.org>

On Saturday, 10 February 2007, at 13:44:03 (-0500),
gregory pryzby wrote:

> Look at how SELinux is still around

There's never been better proof that shitty code for unreliable
systems of self-defeating complexity can still make it into the
mainline Linux kernel.

Just Say No to ShitELinux.

Michael

-- 
Michael Jennings (a.k.a. KainX)  http://www.kainx.org/  <mej@kainx.org>
n + 1, Inc., http://www.nplus1.net/       Author, Eterm (www.eterm.org)
-----------------------------------------------------------------------
 "I've never seen you looking so gorgeous as you did tonight.  I've
  never seen you shine so bright; you were amazing.  I've never seen
  so many people want to be there by your side, and when you turned to
  me and smiled, it took my breath away."
                                   -- Chris Deburgh, "The Lady in Red"

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: chris@dibona.com
From: Theodore Tso <tytso@mit.edu>
Cc: evals@lists.merlins.org
Date: Sat, 10 Feb 2007 14:05:14 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>
Message-Id: <20070210190514.GH25607@thunk.org>
References: <1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209132255.GA9902@thunk.org>
	<20070209132950.GA30104@pryzby.org>
	<20070210163649.GD25607@thunk.org>
	<7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>

On Sat, Feb 10, 2007 at 10:30:54AM -0800, Chris DiBona wrote:
> It would seem somewhat unethical to compete with your former company in
> which you would have a founder level of stock. Maybe he is a fan of
> virtualization? It's no secret that Xen went through a , ahem, reimagining
> last year.

I agree it would certainly be unethical if he still owns a founder
level of stock in Xensource

One additional interesting data point.  On January 14, 2007, someone
(anonymously, from an IP address that appears to be from an ADSL
account in the Netherlands) added the following sentence to Moshe
Bar's Wikipedia entry:

   Moshe is also founder and CTO of the company Qumranet which is behind
   the development of the KVM virtualization technology in the Linux
   kernel.

Now as we all know, information in Wikipedia isn't necessary accurate,
and I have some issues with the second half of that sentence, since
there is lot of KVM being developed by other people (such as Ingo
Molnar at Red Hat).  However, in combination with the fact that the
KVM wiki is hosted at Qumranet, it certainly makes you go hmmm...

						- Ted

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: "Theodore Tso" <tytso@mit.edu>
From: "Chris DiBona" <cdibona@gmail.com>
Cc: evals@lists.merlins.org
Date: Sat, 10 Feb 2007 11:17:14 -0800
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070210190514.GH25607@thunk.org>
Message-Id: <7d9492d90702101117m487c89c0nbb2e9e467d658424@mail.gmail.com>
References: <1170951636.3916.13.camel@localhost.localdomain>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org> <20070209132255.GA9902@thunk.org>
	<20070209132950.GA30104@pryzby.org> <20070210163649.GD25607@thunk.org>
	<7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>
	<20070210190514.GH25607@thunk.org>

It makes me go More than hmm, it makes me think that if you have
Moshe/Qumranet work on KVM then you might be exposing linux to unwanted
intellectual property issues. It makes me think that Moshe is using
linux/open source as a way of getting around his non-compete with Xen. Mind
you, I'm assuming that he signed a founders style release of ip and
non-compete which can be way more draconian and enforceable than a standard
employee release. I'm probably being too conservative as well. I'd rather
KVM takes longer than introduce these kinds of issues into kvm.

Chris

On 2/10/07, Theodore Tso <tytso@mit.edu> wrote:
>
> On Sat, Feb 10, 2007 at 10:30:54AM -0800, Chris DiBona wrote:
> > It would seem somewhat unethical to compete with your former company in
> > which you would have a founder level of stock. Maybe he is a fan of
> > virtualization? It's no secret that Xen went through a , ahem,
> reimagining
> > last year.
>
> I agree it would certainly be unethical if he still owns a founder
> level of stock in Xensource
>
> One additional interesting data point.  On January 14, 2007, someone
> (anonymously, from an IP address that appears to be from an ADSL
> account in the Netherlands) added the following sentence to Moshe
> Bar's Wikipedia entry:
>
>    Moshe is also founder and CTO of the company Qumranet which is behind
>    the development of the KVM virtualization technology in the Linux
>    kernel.
>
> Now as we all know, information in Wikipedia isn't necessary accurate,
> and I have some issues with the second half of that sentence, since
> there is lot of KVM being developed by other people (such as Ingo
> Molnar at Red Hat).  However, in combination with the fact that the
> KVM wiki is hosted at Qumranet, it certainly makes you go hmmm...
>
>                                                 - Ted
>



-- 
Open Source Programs Manager, Google Inc.
Google's Open Source program can be found at http://code.google.com
Personal Weblog: http://dibona.com
_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals

To: chris@dibona.com
From: Theodore Tso <tytso@mit.edu>
Cc: evals@lists.merlins.org
Date: Sat, 10 Feb 2007 14:29:15 -0500
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.5.12-2006-07-14
List-Id: <evals.lists.merlins.org>
In-Reply-To: <7d9492d90702101117m487c89c0nbb2e9e467d658424@mail.gmail.com>
Message-Id: <20070210192915.GI25607@thunk.org>
References: <20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209132255.GA9902@thunk.org>
	<20070209132950.GA30104@pryzby.org>
	<20070210163649.GD25607@thunk.org>
	<7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>
	<20070210190514.GH25607@thunk.org>
	<7d9492d90702101117m487c89c0nbb2e9e467d658424@mail.gmail.com>

On Sat, Feb 10, 2007 at 11:17:14AM -0800, Chris DiBona wrote:
> It makes me go More than hmm, it makes me think that if you have
> Moshe/Qumranet work on KVM then you might be exposing linux to unwanted
> intellectual property issues. It makes me think that Moshe is using
> linux/open source as a way of getting around his non-compete with Xen. Mind
> you, I'm assuming that he signed a founders style release of ip and
> non-compete which can be way more draconian and enforceable than a standard
> employee release. I'm probably being too conservative as well. I'd rather
> KVM takes longer than introduce these kinds of issues into kvm.

But most of the low-level virtualization in Xen is open source and/or
public information (i.e., research papers written by Ian Pratt and the
rest of his research group at Cambridge).  The stuff which is
Xensource's secret sauce was all of their management layer on top of
the Xen open source code.  

Given that KVM was written from scratch by others[1] not using any of
the original Xen para-virtualization approach, and it's other people
who as far as I know are not associated with Qumranet who are thinking
about taking Rusty's paravirt_ops (developed by IBM engineers so that
the same virtualization interfaces could be used by both VMware and
Xen) and plugging KVM into it, it's not at all clear how there could
be any patent, copyright, or trade secret entangelements in the core
KVM code.

There may be some non-compete issues, but those are contract issues
between Xen and Moshe, and wouldn't be IP issues that might require
that the KVM code to be yanked out of the Linux kernel --- and even if
it did, it's all under git these days so it would be trivially easy to
pull it out if it came to that.

I could potentially image some IP issues in the management layer code
in Qumranet --- assuming that their business model involves competing
head-to-head with XenSource --- which we don't know at this point;
only that it seems likely that Qumaranet is doing _something_ with the
KVM technology.  But that's not really a concern of the Linux kernel.

Still, you raise a good potential point.  Perhaps we should ask
someone like the Software Freedom Law Center to render a quiet legal
opinion to Linus and Andrew...

						- Ted

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals


To: "Theodore Tso" <tytso@mit.edu>
From: "Chris DiBona" <cdibona@gmail.com>
Cc: evals@lists.merlins.org
Date: Sat, 10 Feb 2007 11:43:38 -0800
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070210192915.GI25607@thunk.org>
Message-Id: <7d9492d90702101143p45d80792g1948786bbba590b6@mail.gmail.com>
References: <20070209005801.GB2847@verge.net.au>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org> <20070209132255.GA9902@thunk.org>
	<20070209132950.GA30104@pryzby.org> <20070210163649.GD25607@thunk.org>
	<7d9492d90702101030x1ef1887aj8b7498d0e8132753@mail.gmail.com>
	<20070210190514.GH25607@thunk.org>
	<7d9492d90702101117m487c89c0nbb2e9e467d658424@mail.gmail.com>
	<20070210192915.GI25607@thunk.org>

That makes a lot of sense, but yeah, this is the kind of thing that Eben is
for. :-)

Chris

On 2/10/07, Theodore Tso <tytso@mit.edu> wrote:
>
> On Sat, Feb 10, 2007 at 11:17:14AM -0800, Chris DiBona wrote:
> > It makes me go More than hmm, it makes me think that if you have
> > Moshe/Qumranet work on KVM then you might be exposing linux to unwanted
> > intellectual property issues. It makes me think that Moshe is using
> > linux/open source as a way of getting around his non-compete with Xen.
> Mind
> > you, I'm assuming that he signed a founders style release of ip and
> > non-compete which can be way more draconian and enforceable than a
> standard
> > employee release. I'm probably being too conservative as well. I'd
> rather
> > KVM takes longer than introduce these kinds of issues into kvm.
>
> But most of the low-level virtualization in Xen is open source and/or
> public information (i.e., research papers written by Ian Pratt and the
> rest of his research group at Cambridge).  The stuff which is
> Xensource's secret sauce was all of their management layer on top of
> the Xen open source code.
>
> Given that KVM was written from scratch by others[1] not using any of
> the original Xen para-virtualization approach, and it's other people
> who as far as I know are not associated with Qumranet who are thinking
> about taking Rusty's paravirt_ops (developed by IBM engineers so that
> the same virtualization interfaces could be used by both VMware and
> Xen) and plugging KVM into it, it's not at all clear how there could
> be any patent, copyright, or trade secret entangelements in the core
> KVM code.
>
> There may be some non-compete issues, but those are contract issues
> between Xen and Moshe, and wouldn't be IP issues that might require
> that the KVM code to be yanked out of the Linux kernel --- and even if
> it did, it's all under git these days so it would be trivially easy to
> pull it out if it came to that.
>
> I could potentially image some IP issues in the management layer code
> in Qumranet --- assuming that their business model involves competing
> head-to-head with XenSource --- which we don't know at this point;
> only that it seems likely that Qumaranet is doing _something_ with the
> KVM technology.  But that's not really a concern of the Linux kernel.
>
> Still, you raise a good potential point.  Perhaps we should ask
> someone like the Software Freedom Law Center to render a quiet legal
> opinion to Linus and Andrew...
>
>                                                 - Ted
>



-- 
Open Source Programs Manager, Google Inc.
Google's Open Source program can be found at http://code.google.com
Personal Weblog: http://dibona.com
_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals

To: Theodore Tso <tytso@mit.edu>
From: n0ano@n0ano.com
Cc: evals@lists.merlins.org
Date: Fri, 9 Feb 2007 20:47:18 -0700
Subject: Re: [Evals] Zimbra: Am in San Mateo this weekend
User-Agent: Mutt/1.4.2.1i
List-Id: <evals.lists.merlins.org>
In-Reply-To: <20070209132255.GA9902@thunk.org>
Message-Id: <20070210034718.GA15734@xolotl.n0ano.com>
References: <5a433ca90702080229l4e07a164mee3093ed3bb3f120@mail.gmail.com>
	<1170950089.3916.9.camel@localhost.localdomain>
	<20070208160454.GB27071@pryzby.org>
	<1170951636.3916.13.camel@localhost.localdomain>
	<20070208163907.GD27071@pryzby.org>
	<20070209005801.GB2847@verge.net.au>
	<20070209023238.GE13207@thunk.org>
	<20070209032731.GA5709@verge.net.au>
	<20070209034512.GI28750@pryzby.org>
	<20070209132255.GA9902@thunk.org>

Without going into the Xen vs. KVM argument (since I know nothing about
KVM) I want to clear up one misconception in your email.  Xen is running
unmodifed Windows quite happily and has been for about a year now.  It
requires a VT machine but I've got it running in my basement right now
so it really does work.

On Fri, Feb 09, 2007 at 08:22:55AM -0500, Theodore Tso wrote:
> .
> .
> .
>
> accelerator, but instead, it's going to be the other way around; KVM
> will be able to run unmodified legacy OS's just fine (which Xen can
> not), as well as running Linux guest OS's very quickly with
> paravirtualization enabled.

-- 
Don Dugger
"Censeo Toto nos in Kansa esse decisse." - D. Gale
n0ano@n0ano.com
Ph: 303/443-3786

_______________________________________________
Evals mailing list
Evals@lists.merlins.org
http://lists.merlins.org/lists/listinfo/evals