[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [ebxml-bp] Manual Operations in BPSS"
Dear Martin Duane and others, Yes, I agree with this - it is exactly the point I was making (and I also acknowledge the other helpful mails from Duane and others). To synchronise state one must not only set the communicating protocol machines to complementary states (which is what we tend to concentrate on), but also align the shared resources (or objects). And yes I picked up on this from a number of folks including Bob Haugen, it is my understanding that it is indeed the central idea of the COOL architecture, but it actually goes back to CMIP as Martin mentions, and probably way before that too. In spite of the COOL presentation I am not sure that the ideas have been properly worked into the UN/CEFACT documents yet though(?) But the alignment of protocol state and resource / object state certainly should be part of BPSS as it goes forward (in my opinion). Best Regards Tony A M Fletcher Home: 35, Wimborne Avenue, IPSWICH IP3 8QW Tel: +44 (0) 1473 729537 Mobile: +44 (0) 7801 948219 amfletcher@iee.org <mailto:amfletcher@iee.org> (also tony.fletcher@talk21.com <mailto:tony.fletcher@talk21.com> & tony_fletcher@btopenworld.com) -----Original Message----- From: martin.me.roberts@bt.com [mailto:martin.me.roberts@bt.com] Sent: 20 February 2004 11:00 To: dnickull@adobe.com; tony_fletcher@btopenworld.com Cc: david@drrw.info; zbarch@rcn.com; ebxml-bp@lists.oasis-open.org Subject: RE: [ebxml-bp] Manual Operations in BPSS" My understanding of the original COOL architecture presented at the ebtwg in Seattle was that the state was held in objects, and that the synchronising of object state is what B2B would be about. In this case if an object is updated manually the state would be sent irrespective of the transaction model being supported. In some ways this is a useful model and creates and shared object model. The telco world used this model with the invention of CMIP which was a protocol that has at its heart a set of managed objects that can be synchronized over the wire when ever they change. It has useful concepts of filters and scopes that allow the partners to come to some agreements as to what attributes to synchronize within each object. The advantage of this approach is that it is not transactional but can be used to indicate transactions, e.g. I create an order on my system. The public part gets sent to my partner who responds with the necessary data and we have a transaction. The fact that there may be time delays etc is reflected by the underlying protocol. It also allows for a partner to work all day updating their system offline and to come online when the system load is light and expose all the objects that have changed and the other systems are updated appropriately. This is where the UN CEFACT model is going I understand. With the loose coupling of web services and the extra resources available to computers today this type of system could work. Bandwidth could well be a problem still although the effects of a 37 baud line really hampered our early implementation which where trying to operate a follow-the-sun scenario. Martin Roberts xml designer, BT Exact e-mail: martin.me.roberts@bt.com tel: +44(0) 1473 609785 clickdial fax: +44(0) 1473 609834 Intranet Site :http://twiki.btlabs.bt.co.uk/twiki
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]