I don't know what's going on since I seem to get these
emails out of order somehow, and there's often a long
delay between when I post and actually receive a
message.  This reply appears to be to a message I
haven't yet seen.

In any case I did want to respond to the CORBA part of
this reply.

It is true that most of our experience to date with
large scale SOA implementations is based on CORBA.  We
are seeing some large scale SOAs based on Web services
starting to emerge, and they are different in
significant ways.  It's possible we are not doing
enough to abstract and generalize the lessons learned
and best practices from our CORBA experiences.

But I think we do recognize that SOA is distinct from
its implementation in technolgy, and while it's true
that many SOAs are based on CORBA, the use of CORBA
per se does not equate to an SOA.

Eric

--- Gregg Wonderly <[EMAIL PROTECTED]> wrote:

> Jan Algermissen wrote:
> > [1] Though I wonder if that is really possible,
> since AFAIK  
> > distributed systems usually need some sort of
> (eventually)  
> > centralized name lookup service and from a
> manageability POV, there  
> > should be some centralized form of 'credentials'
> management (e.g.  
> > NIS, LDAP) so they need not be spread across all
> nodes of a given  
> > system.
> 
> I think there is a big difference between a "central
> service" and a "centrally 
> important service".  I keep sensing a lot of CORBA
> ORB oriented references in 
> all of this SOA vs Object Systems.  I think there is
> an attempt at saying, 
> indirectly that SOA's shouldn't contain an ORB
> because that was a bad experience 
> for some (if not many) who didn't know how to build
> a distributed system.
> 
> But, I can only guess at these things, since I'm not
> them.
> 
> Sure are some very interesting and seemingly oddly
> focused comments about what 
> SOA really is floating around.
> 
> An SOA to me is an architecture where you have the
> freedom to add a separate 
> service to solve a problem just as easily as you
> might add a new function/method 
> to an existing API.  It's the software system that
> enables you to do this.  If 
> your system architecture is simply constructed as a
> composite, single process 
> system, where no parts can be separated to a second
> system to scale CPU or 
> distributed geographically to manage long distance
> latency or connectivity 
> failures, it's not SOA.
> 
> If your software development platform includes APIs
> that take the 8 fallacies of 
> distributed computing into account and support a
> software architecture and 
> associated APIs that encourage remoting of
> operations at whatever granularity 
> make sense, then you have the ability to create an
> SOA with that platform.
> 
> But, you still have to do the right things to make
> SOA a reality.
> 
> Gregg Wonderly
> 
> 
> 
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 




 
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/
 



Reply via email to