It probably hasn’t escaped your attention that my blog frequency has dropped off lately. One reason is the insane travel schedule. Another is trying to spend time with my boy, rather than staying late in the office to write that “one last blog”.
Nigel Fortlage sent me an email asking what I think of the Microsoft Novell agreement, with respect to his own thoughts. Who I am not to respond to my community, especially when the request comes from a guy that is walking the talk: running OSS in production environments on IBM System I gear, for all core business functions. He is a SUSE customer, in other words, but errs on the side of open distros and tools. He is a cheap technology officer….
First off I should point out that my esteemed colleague, and all round open source guru, Stephen, has already covered the news in depth, which is good because it means I can focus on the bright and shiny stuff.
Two Heads Better Than One?
First off I should say the deal is a good thing for enterprises and service providers, which invariably just want to be able to deploy technologies that work together. Some people think Novell shouldn’t work with Microsoft. Some people think open source projects shouldn’t talk to or be willing to work with Microsoft. But I think a mutually exclusive approach is asinine. When Microsoft called to brief me on the news I believe I said:
“This is good news. We’re not supposed to be in a sandal-wearing pissing match.”
The more closely Microsoft and Novell work together the better. There is nothing to stop Red Hat from joining the party, but it would need to sign a licensing agreement of its own. Surely some of the groundwork could have been laid with the JBoss Microsoft agreement?
Whether or not RedMonk is open source and open standards oriented (which it is), we maintain an enterprise focus in our work. I believe the news needs to be examined in that context. Anything that helps with enterprise interop is to be applauded, and from what I hear this deal involves a lot more engineering than similar announcements between for example Sun and Microsoft. Vitalization and identity would seem to be key to the deal.
Attack on The Clones
My biggest concern though is the cut off date for the agreement – 2012. At that time, according to Bill Hilf (pictured right), who I was lucky enough to meet yesterday (the charmed life of an analyst/blogger), the covenant not to sue will be reviewed at that time.
If, and this is the scary bit, Microsoft considers its technology has been “cloned” at that time, then things could get really ugly. The covenant not to sue is not irrevocable, which is still chilling, if no longer downright icy. Stephen has written on IP and Mono before. In case you’re wondering what the hell Mono is, Stephen’s report here explains.
Lets consider a scenario where grassroots developers and ISVs have used Mono, the open source implementation of the C# and Common Language Runtime (CLR) ECMA standard, to build really cool apps, which become widely deployed. Then – Microsoft decides to activate its “non-cloning” strategy. What then is the only logical choice for people writing, maintaining and using these applications? Why an easy port to Windows… Of course anti-trust law might trump IP in this case, but that’s a lawyer’s question. One other opportunity for Mono refugees in 2012 might be to migrate the other way- from .NET to JEE, using Mainsoft‘S rather spiffy code software, which it positions as VisualStudio for JEE. IBM should probably buy Mainsoft, because of its role in potentially bridging .NET and Java… it could do with some control points in this area. But of course if Microsoft wanted to shut Mainsoft down this could be a moot point.
So while I am positive on the news, I am maintaining a skeptical position. We need to encourage Microsoft to do more though by saying well done folks, rather than throw stones for its own sake.
Psyche! Clone.
I think Microsoft needs to drop its obsession with clones.
IBM was forced to support clones of its mainframes through the consent decree, which saw the rise of plug compatible manufacturers such as Amdahl, HDS and Comparex. How did IBM win in the long run? By out-engineering its competitors, not through the use of IP law. Competing on the basis of implementation and brand is the only really honest competition. I still believe that the antitrust actions against IBM show us what can be done to level playing fields, and reduce lock in and market dominant behaviour.
Mono is an implementation of a standard, not a clone. If Microsoft didn’t want to encourage cloning it should have never gone to ECMA in the first place. But then I am not a lawyer. David Berlind has some good coverage of the patent and indemnification related issues here, particularly as they relate to Red Hat. That’s a company taking major hits at the moment. I expect to see RedHat start making some more acquisitions commercial open source space in the near term, to buy more IP into the company.
Virtualisation Cracks
When Jim Ni, group product manager in Microsoft’s virtualisation group, briefed UK analysts a couple of months ago he took a position which I felt, and still feel, is somewhat absurd. He said that in just a few years we might see virtualisation bifurcate into two key technologies – Xen and Microsoft Virtual Services. VMWare, he contends, will hit the skids. Or as Hilf put it succinctly yesterday:
“Unless we screw up on execution we’ll just gobble up their market”…
Wow. That antitrust training really knocked the stuffing out of you, didn’t it Bill?
I personally think all of VMWare’s competitors underestimate the power of the franchise – especially when you consider for example that financial services companies in the City of London are increasingly mandating not Linux skills or Microsoft skills, but VMware skills. What is more, VMware is an innovator, and keeps moving fast – the virtual appliance idea is another doozer.
But, with the news that Novell and Xen are going to be able to work closely with Microsoft in the virtualisation space does begin to create some interesting new dynamics, that I am sure Red Hat and VMware won’t be too enthusiastic about.
What else? Well if SUSE will run Windows then that means the IBM mainframe should be able to run Windows server as a supported environment. Support is key to the deal. Thus for the longest time (until it was ready with its own virtualisation technology…) Microsoft refused to support VMware as a production environment for applications. Here though the is and ts are already dotted and slashed: Novell customers are going to be able to run Windows apps on their servers, and vice-versa, fully supported by both vendors. That is enterprisey goodness.
Identity Odd Man Out
Identity feels like the odd man out in the agreement. After all, Novell has a ton of IP in this area, which doesn’t rely on open source. I am pretty sure Novell would come up smiling in a patent shoot out with Microsoft around identity and directory. From a customer perspective however, more effective interop between Active Directory and eDirectory has to be a good thing. Interesting Sun has made no announcements in this space since last week…
Piece Out
This blog was never intended to be comprehensive. Indeed-there is a lot in the news that will shake out over time. But I put my 2c worth in – that work for you, Nigel?
disclaimers: Stephen is a long time Mono fan but for some reason Novell is not a customer. Microsoft does occasionally pay us money, but its not really a RedMonk patron. Stephen is at a VMware conference now, and I look forward to coverage, but its not a client, and neither is Xen.
tags:
xen, VMWare, Microsoft, virtualisation, Red Hat, IBM mainframe, Novell, eDirectory, ECMA, Mono
Nigel Fortlage says:
November 8, 2006 at 2:24 am
James, well done. WOW what a lot of travel… I am sure all family members are glad to have you home now.
Any passing thoughts on the open document standards part of the Microsoft/Novell deal…I see a lot of positive potential for that, especially since Novell has committed the end results back to openoffice.org…