WS Reliable Messaging creeps forward
Two standards, one year, little progress
One year after BEA, IBM, Microsoft and Tibco (BIMT) published a draft specification for WS Reliable Messaging - hot on the heels of a rival spec from an Oracle, Sun, Sonic, Fujitsu, Hitachi and NEC consortium - the question is this: has any progress actually been made, asks Bloor Research analyst Peter Abrahams.
WS Reliable Messaging describes a protocol that allows messages to be delivered reliably between distributed applications in the presence of software component, system, or network failures. Its rival - WS Reliability - addresses the same problem. The latter was published in January, offered to OASIS in February, and now faces the scrutiny of a new technical committee.
This month, BIMT published an updated version of its work. Amendments are based on the feedback from a workshop last July and a further interoperability workshop in October.
So, twelve months on, what's new? Unfortunately, it's difficult to say because there is no document that lists the changes. The difference between original and amended specs seems negligible.
Some of the terminology has changed: for example in 2003 the original sender of a message was called 'RM Source'; in 2004 it has become 'Sender'. This is hardly earth-shattering, but probably reflects a level of generalisation which will allow different messaging protocols to work together. That aside, the general architecture and message flows seem to be identical.
So, has the wait been worth it? Well, the validation was useful and the interoperability testing proved that the specification is robust. However, the fact remains that there are still two competing specifications. The intervening year has simply seen both even more firmly entrenched.
The competing specifications are not hugely different. In reality, they share the same provenance, and it would hugely benefit the industry as a whole if they came together. BIMT includes IBM and Microsoft so they have the commercial clout to make their specification the de facto standard. Nevertheless, BIMT have yet to say when, or even if, they will propose the specification to a standards body.
It's imperative that BIMT do this as quickly as possible so that a single standard can be agreed. The great strides that have been made because of the base web services standards (SOAP, WSDL, XML and UDDI), show that single standards are good for everyone in the medium term and conflicting standards would be disastrous in the long term.
Sponsored: Hyper-scale data management