* Liberty chief calls for converging of Liberty and WSI specs
After I castigated the Liberty Alliance for releasing its white paper “Liberty Alliance & WS-Federation: A Comparative Overview” during the recent Digital ID World conference, I expected to hear from Liberty President Michael Barrett. I wasn’t disappointed.
Within hours of the newsletter’s release, Barrett’s reaction was in my in-box. Within a couple of days, we were “mouth-to-ear” in a telephone conference. While I’m still not convinced that the paper was necessary, or even beneficial, I’ll present Barrett’s argument and let you decide for yourself.
According to Barrett, one of the major reasons why identity federation projects are moving slowly is the uncertainty of businesses as to which federation standard to support.
Businesses would like to see a single standard that everyone can incorporate into their federation products and technologies. The paper, therefore, was to show the overlaps and differences between the Liberty Alliance specification and the WS-Federation specification, which is offered by the IBM-Microsoft-BEA consortium as part of its Web Services Initiative (WSI). The “call to action” in the paper was for an all-parties conference to begin the process of converging the two specs into a single initiative.
You certainly can’t argue with that objective: We’d all like to see a single specification for any technological initiative. Unfortunately, that’s rarely how it works in the real world.
Very few single standards can satisfy everyone’s needs and desires. But sometimes, in the spirit of cooperation, we need to compromise our wants for the good of society as a whole. The reason is usually because competing standards are mutually exclusive or would be prohibitively expensive to interoperate. But neither of those conditions exist with identity federation.
I asked a number of the technologists involved in the various federation standards-setting arenas how difficult it would be to support both the Liberty specification and the WS-Federation specification within a single application or service. No one said it couldn’t be done, and most said it was nowhere near being a showstopper.
If the non-technical members of the Liberty Alliance need to be educated as to the ability for different standards to interoperate, that’s a useful exercise. Letting them drag their feet on establishing identity projects while they wait for convergence that may never come, though, is a sure way to spin everyone’s wheels and ensure that the Liberty specification (which from a business view is far superior to WS-Federation) languishes.
The technology landscape is littered with proposals, specifications and standards that were never finalized, tested or implemented. Often these were technically superior to the proposals, specs and standards that were actually implemented. But technical superiority has never been a formula (or even a pre-requisite) for commercial success.
Come on, Liberty, let’s see an actual, functioning, commercially-viable “Circles of Trust” (the basic building block of a federated identity group in the Liberty specification). That, more than any paper or pleading, could drive the WSI people to the bargaining table.
