|
There's nothing much I can say about Sun. On the Microsoft side, they've published their position paper regarding ESBs in the past. To me, the whole message seems consistent with an predisposition to Microsoft-everywhere. Clearly, if you have the same platform everywhere, there's no need for an ESB. That position is also consistent with a smart nodes, dumb network approach. The same thing would hold true if everything was WebSphere or Oracle. Microsoft's probably the only company that can actually expect to have an all Windows environment, however.
Microsoft is at least very consistent with its message. I have no issues with service-oriented computing. After all, Microsoft is selling the computing servers and development tools behind the services, not the services themselves. We all know you can't buy an SOA. If anything, we'd all be better off if SOA became more of a business trend than a technology trend. Service oriented computing would be what IT does in support of the business SOA.
-tb On Mar 6, 2006, at 4:21 PM, Anne Thomas Manes wrote: I'll provide a slightly different perspective.
On Microsoft:
Microsoft defines its own path and doesn't like to follow the crowd. Microsoft views SOA as too grandiose, and it is instead focusing on more practical, pragmatic, and tactical issues rather than life-style changing, strategic initiatives like SOA.
The Microsoft Servers and Tools group (responsible for Windows Server, Windows Server System, .NET, and Visual Studio) has repeated told me that they aren't comfortable with service-oriented "architecture". Instead they are focusing on service-oriented "computing". The buzz word for their marketing message is "Connected Systems", not SOA.
This is the group that focuses on the enterprise software market: a superplatform that competes with the likes of WebSphere, WebLogic, and Oracle Fusion. I agree with Dan that this group does not view an ESB as an end game opportunity. In fact, this group dismisses an ESB as, at best, a temporarily marginally useful middleware product that fills in a few gaps while the WS-* specs are being finalized, but will soon become irrelevant. I seriously doubt that this group will ever produce an ESB. It already has BizTalk, and it's about to release Windows Communication Foundation (WCF, aka "Indigo").
Meanwhile, Microsoft is focusing on making it easier to build new apps and to integrate old ones. .NET and WinFX (aka "Indigo", "Avalon", and Windows Workflow Foundation) are excellent frameworks for rapidly building applications and making applications communicate. Reusable services, business alignment, and other terminology often associated with SOA are not part of Microsoft's Connected Systems vocabulary.
The more strategic part of Microsoft's Connected Systems message has to do with designing systems for operations -- the Dynamic Systems Initiative (DSI) and System Definition Model (SDM) stuff.
Other groups within Microsoft are focused on productivity applications (Office) and the consumer market. As Dan said, these groups are very interested in the SaaS business model, as evidenced by the Office Live and Windows Live initiatives. But these efforts are orthogonal to the superplatform group efforts.
On Sun:
Dan nailed it. Sun doesn't know how to market or sell software, and it's not willing to invest in retraining the sales force. Sun is a hardware company. And the only software that it can possibly sell is software that's a proprietary solution or packaged with hardware.
Consider this: Sun acquired SeeBeyond in an effort to get in on the SOA opportunity. How completely typical that they couldn't comprehend the difference between a proprietary integration technology and standards-based approach to SOA.
Pitiful.
Sun has developed one product that is reasonable well targeted at SOA -- it's new registry/repository system. Although this product isn't sold separately. It's only supplied as part of the Sun Java Enterprise System.
Anne
On 3/6/06, Gervas Douglas <[EMAIL PROTECTED]> wrote: Daniel,
What your write about MS and their SaaS plans fits in with statements
I have read. Even so, not everyone will leap unhesitatingly at this
reinvention of the computer bureau (can you imagine GCHQ rushing into
this, to give an extreme example?), and that means that large
organisations now and later are going to demand to know more about
what they offer on the SOA front.
As regards Sun, I could not possibly expand upon what you have stated
as I have not had anything like the extent of your direct contact with
them (Team, Dan is one of the few people on this Group whom I know,
although it is a couple of years since we viewed each other's ugly
mug). I did once have a drink with the great Scott McNealy and I
found him charming and tolerant (he needed to be the latter at that
stage of the evening). We did not discuss SOA as this was in 1992.
No, before anyone asks, I have not had the honour of meeting Gregg.
However it would be very interesting to hear from some Solar members
of this Group how Sun's strategy will take advantage of their
interesting software assets.
gsd
--- In [email protected], "creswell_dan"
<[EMAIL PROTECTED]> wrote:
>
> On MS's technology fit to SOA:
>
> The problem with SOA is that everyone and his dog has what they deem
> an SOA product. Interestingly, in most cases, the product they offer
> is very similar to what they were offering prior to the "rise" of SOA.
>
> MS's approach seems to be much more in the vein of assumptions that
> the SalesForce.com model is one that most people will adopt. I think
> they are working at a story there but that's all about external facing
> services. I've seen less about internal facing (i.e. behind the
> firewall) services. I suspect that MS would say the technology is
> suitable for both kinds of service but I sense that their emphasis is
> being directed by their expectation of where the big bucks will be made.
>
> And if MS are being coy as you put it, I think that's because they see
> the end game being not ESB or SOA itself but the external service's
> market as exhibited by SalesForce.com and so they're trying to leap
> over a couple of intermediary steps thus gaining a lead. They'll
> backfill to ESB's etc if they need to but they are after a much more
> valuable beachhead.
>
> On Sun and their software:
>
> I think it's horrifyingly simple - Sun still don't see any value in
> software other than as a means for getting you to buy their
> hardware/OS. Schwarz has recently been making overtures to HP about
> merging stuff with Solaris 10. That could definitely be interpreted
> as "HP and Sun are getting commoditized out of the market, we need to
> be bigger, let's join forces and customer bases". Smells like
> consolidation to me.
>
> So why are Sun coy? They're not - they're naive/inept - they don't
> even know they have software that plays in the area of SOA! Being
> specific, there are various engineers who do know that but the sales
> side of the organization can't see it, and can't guarentee to sell
> boxes with it ignoring the fact that they could sell lots of software
> boxes. Note also that all indications I have suggest that most of the
> engineers don't get it either - they still inhabit a three-tier or
> client-server world and can't/won't grok services and are still
> building much of their software to function in that old world.
>
> Two cents from a techie,
>
> Dan.
>
> --- In [email protected], "Gervas
> Douglas" < gervas.douglas@> wrote:
> >
> > We read quite a lot in this Group about Java and other non-MS
> > technologies. And this depite Sun not being over-aggressive about
> > marketing software, and doing even less for some of its Orphan Java
> > technologies (how often do they bother to mention Jini, RIO, Jxta
> etc.??).
> >
> > What we don't read or hear much about is what is going on in the
> > Microsoft SOA universe. I know MS do not do much to push the concept
> > of SOA, but a lot of SOA implementations take place in .NET
> > environments. Most big organisations seem to have a mixture of .NET
> > and Java. So what are MS doing about SOA middleware now that every
> > platform vendor and his dog is offering an ESB? There is of course
> > the spectre of Indigo on the horizon, but it is not readily apparent
> > how this is going to fit into the SOA middleware scene. Do any of you
> > have any information on this? Come to think of it, why are Sun and
> > Microsoft so coy about these key aspects of their technology?
> >
> > Gervas
> >
>
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/service-orientated-architecture/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
SPONSORED LINKS
YAHOO! GROUPS LINKS
SPONSORED LINKS
YAHOO! GROUPS LINKS
|
