RE: [wsrp-wsia] Issue #19 - Hints for InitEnvironment

[Andre Kramer <andre.kramer@eu.citrix.com>]

The proposal was to make it optional for both producers and consumers but,
as you suggest, I agree it could be used to make the producer sharing
explicit (e.g. for JSR168 Application scopes). 

[I'm not sure I understand your registration cloning usage. Was this to get
more than one application grouping out of a producer? Seems a bit costly to
register once application grouping ... I might start asking for Leasing
(controlling registration timeouts) again ;-)
]

regards,
Andre

-----Original Message-----
From: Rich Thompson [mailto:richt2@us.ibm.com]
Sent: 07 October 2002 16:03
To: wsrp-wsia@lists.oasis-open.org
Subject: Re: [wsrp-wsia] Issue #19 - Hints for InitEnvironment







If we add this piece of metadata about how entities could benefit from
sharing, would we also strengthen the meaning of groupID by requiring that
Producers use it to scope the sharing of data? I would prefer this as it
gives real semantics to including any concept of sharing in the protocol.
In addition, this would stop Consumer from having to clone registrations in
order to be able to control data sharing.


 

                      Andre Kramer

                      <andre.kramer@eu.        To:
wsrp-wsia@lists.oasis-open.org                          
                      citrix.com>              cc:

                                               Subject:  [wsrp-wsia] Issue
#19 - Hints for InitEnvironment       
                      10/07/2002 07:07

                      AM

 

 




Since I originally raised #19 and as I've not seen any follow on discussion
on our infamous initEnvironment()/groupID topic since last week's call, I
would like to make more concrete a proposal made in the call's lengthy
groupID discussion.

7.1.1. states: "Since Consumers will not know what entities could benefit
from any Producer mediated data sharing, they SHOULD place all entities
from a Producer within a single group until a reason is identified for
specifying a separate groupID".

I agree this is enough to support JSR168 application/shared session scopes
by providing each user with a separate (i.e. one) load balanced environment
via the groupID argument to initEnvironment(). However, an extra hint from
the producer is required in order to load balance a user's Portlets across
multiple JSR168/Web Applications and nodes:

(Proposal) A producer could include a "String preferredGroupID" (as an
optional hint) in an EntityDescription element (or possibly in the
EntityType meta data). Consumers could then discover this extra performance
related meta data using getEntityDescription() and use it for the groupID
in initEnvironment(), as described in 7.4. I believe they are not required
to do so, but I would still say "SHOULD" rather than "CAN", which is why
this may be more than just a "vendor" extension for JSR168.

Hope this helps us further the discussion,

Andre
      -----Original Message-----
      From: Gil Tayar [mailto:Gil.Tayar@webcollage.com]
      Sent: 07 October 2002 07:30
      To: wsrp-wsia@lists.oasis-open.org
      Subject: [wsrp-wsia] WSRP-WSIA Issues List.xls

      A new version of the issues list.

      Changes:

      1. We now have a manageable number of open issues!

      2. The date whereby a tentative resolution is considered final (due
      to lack of interest) is now in the list, and not one of the secondary
      headers. This will help us browse through the list and see whether
      anything needs to be discussed. BTW, I always set it to two weeks
      from the time I get a tentative resolution from anybody, and always
      send an email on that regard to the list to see whether anybody is
      interested in contesting that resolution.

      Cheers,
      Gil Tayar
      WebCollage





----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>