Re: [nm-wg] immediate future of current schemas

[Warren Matthews <warren.matthews@oit.gatech.edu>]

Hi Tanya

I agree although I guess it is possible that similar tests are done at
similar times. A unique ID would make sure the requester picks up the
results from the test that they requested. I acknowledge that is a
flimsy excuse. 

On the other hand, does it hurt to assign a unique ID? It could be
optional when requesting the result. This should work as long as the
requested test result is put in the database along with everything else.

I will want to keep track of who is making the requests, but that is
probably a separate issue.

I would very much like to say "come back later". Having the client hang
for possibly minutes seems wrong. Is an identifier your only objection?




On Wed, 2004-10-20 at 13:37, Tanya Brethour wrote:
> > 4. Discuss the inclusion of a field for transaction IDs in both the request
> > and existing publication
> > schema [1], to allow better tracking of interactions (ties in wirh whether
> > we want stateful or stateless batching of results).
> 
> I think including any state into the schema is a bad idea. I believe the
> discussion behind this was so that in the case where results are not
> immediately available.. the requester gets and ID that is then used later
> to retrieve the results. 
> 
> What is wrong with the response just saying "come back later" and the
> requester is responsible for coming back and retrieving the results (using
> same request as before, but by the time it comes back the tests will have
> finished)? That way you don't have to worry about IDs being unique or
> storing the request.
> 
> Maybe there is something I am missing that makes an ID necessary.
> 
> I also have a few things to add to the list:
> 
> - HopList: we need to figure out how we want to represent this
> 
> - AvailableBandwidth: Right now.. I don't see anything in the schema that
> actually stores the available bandwidth value.
> 
> - There is an error in the current schema that still has not gotten fixed.
> 'Host' is never used.
> 
> - Lastly, if Dan and Martin are busy with the new schema, we really need
> to get the current response schema (and request) in a cvs like system with
> access to some other people within the NMWG.. so that the changes can be
> made and the website is updated. 
> 
> Thanks,
> Tanya
> 
>