OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

wsdm message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [wsdm] Terminology in the requirements


Muhamed,

Here is a little background on this subject, since you are new to the TC.

1.  The WSDM TC started in part by going through a briefing by co-Chair 
Heather Kreger (Shows up as E1014.ppt in the Reference Material Folder - 
it is older than 30 days, so it doesn't show up by default).  In there, 
the notion as expressed on slide 31, "Managers always ‘talk’ to the 
resource while the actual Port may be any number of agents".  Slide 28 
shows the difference in the WSDL - basically the management port URI 
points to another system or application than the business port URI does 
for an agent situation.

Note that for MUWS of manageable resources that are not Web Services, 
there would only be WSDL for management, making it very difficult, if 
not impossible to figure out if there is an agent involved.

Also, since Web Services can be highly distributed, it is also less 
clear what even constitutes an agent.  Suppose you have 7 processes in 
one WSEE that provide a particular Web Service Endpoint functionality. 
Now you add an 8th process dedicated to management.  Is that 8th process 
just another part of the Endpoint or is it really an Agent?

2.  The MPTC (now dissolved, previous to the WSDM) had a "Service Access 
  Point" for the management interface.  The WSMF submission specifically 
calls out a "managed object" that handles the management for the managed 
resource.

As you point out, though, the only interesting aspect of this is if the 
"manager" or some other entity can make use of knowing whether there is 
an agent that is usefully distinct from the managed resource itself.

There has also been discussion about aggregation of resources and 
whether knowing about a specialized type of agent/manager that is an 
"aggregator" would be useful.

Interestingly, I found Heather's presentation very interesting to review.

Aganagic, Muhamed wrote:

> I am breaking my promise, but I have never implied that agents and 
> managed objects are in a 1-1 relationship. In fact, a single object is 
> typically handled be more then one agent and one agent handles  more 
> than one object. I thought this was understood. If this idea ever gets 
> of the ground one would have to start talking about binding interfaces 
> to agents, agents publishing what interfaces they support and all that jazz…
> 
> That’s actually when things would get interesting and cool.
> 


-- 

John DeCarlo, The MITRE Corporation, My Views Are My Own





[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]