OASIS WS-Transaction (almost) a standard

by Mark Little on Jan 24, 2007 |

The OASIS WS-TX technical committee held a face-to-face meeting last week at IBM Hursely. This is likely the last such meeting prior to final standardisation of WS-Coordination, WS-AtomicTransaction and WS-BusinessActivity. It has been a long struggle to get here, dating back to extended transaction work at the OMG, a first attempt at Web Services standardisation via BTP and OASIS WS-CAF (where there is a lot of overlap due to history). This meeting was primarily just making sure that all of the i's were dotted and t's crossed and getting agreement to progress to OASIS standard; all of the heavy work had been done over the past 12 months.

This is an important step for enterprise Web Services deployments for a number of reasons:

WS-AtomicTransaction defines a traditional ACID transactions model, based on two-phase commit (2PC). Many people believe that 2PC is not right for the loosely coupled nature of Web Services and they'd be right. However, this overlooks the other important aspect of Web Services: interoperability. Interoperability between existing vendor implementations is key for many transaction processing deployments, particularly those that have grown via acquisitions of heterogeneous technologies. This has been a transaction processing holy grail for many years.

WS-BusinessActivity provides a forward compensation model, much more appropriate for loosely coupled, long duration interactions. This latter model will probably have slower take-up than WS-AtomicTransaction, but it should become more important over time.

Once we finalise WS-Security and WS-Reliable Exchange we'll finally be able to do secure, reliable and transacted Web Services in a standard manner.

Related Editorial

Enterprise-class API Patterns for Cloud & Mobile

The Rise of the Composite API Platform, a ‘Best-of-Breed’ API Management Solution

A Unified Mobile Architecture for the Modern Data Center

Speed Matters: See why Nike, Groupon, and Comcast are using New Relic

APM Survival Guide from a Veteran Performance Geek

Hello stranger!

You need to Register an InfoQ account or to post comments. But there's so much more behind being registered.

Get the most out of the InfoQ experience.

Tell us what you think

Community comments

Agree wholeheartedly by John Harby Posted

Unbelievable by Eirik Maus Posted

Re: Unbelievable by Mark Little Posted

Agree wholeheartedly by John Harby

I definitely agree that this is a huge milestone. I recall Sven Frolund, Kannan Govindrajan, myself and others at HP labs discussing the compensation model back in the late '90's. Having participated with Mark and the others on this TC, I would like to state that the TC was conducted with the utmost formality and professionalism and for myself was certainly a learning experience in how such things should be handled.

Unbelievable by Eirik Maus

Have anyone heard of any actually interoperable implementations yet?

Re: Unbelievable by Mark Little

You mean apart from the interoperability work that many of the vendors within the TC did as part of developing the specification? jboss.org/jbossBlog/blog/mlittle/