Craig Burton cries 'ubiquity'

Craig Burton cries 'ubiquity'

Summary: At Novell, Craig Burton was one of the driving forces behind the modern notion of a network as a collection of services rather than a collection of wires.  He's a master at seeing the big picture and identifying the limitations of particular strategies within that picture.

SHARE:
TOPICS: Browser
3

At Novell, Craig Burton was one of the driving forces behind the modern notion of a network as a collection of services rather than a collection of wires.  He's a master at seeing the big picture and identifying the limitations of particular strategies within that picture.  I've known him for years and have great respect for his opinion.

In a recent post on his blog, called I Cry Ubiquity, Craig analyzes current strategies around Internet Identity.   Identity 1.0 was about server-based authentication to services.  Identity 2.0 is about network-based user verified credentials.  Dick Hardt likes to talk about them working just like credentials in the physical world: portable and verifiable without having to constantly talk to some central authority.  Dick's favorite example is using your driver's license to buy beer.  Can you imagine how poorly it would work if the clerk at the convenience store had to call up the Motor Vehicles division every time to make sure the credential was valid?!

As Craig says, Identity 2.0 changes everything.  But, only "when Identity 2.0 infrastructure becomes ubiquitous. Free. A given. Like air and sunshine."  For that to happen, though the infrastructure must meet certain prerequisites.  If you build a proprietary protocol that requires people to adapt to you or not play the game, then you're not likely to achieve the level of ubiquity that Craig's looking for.  Open and simple are necessary, but not sufficient.  What we need is an "that is independent of mandated adoption."  Craig holds up Kim Cameron's Identity Metasystem as an example of a system that meets this requirement. 

Kim, who is Microsoft's Chief Identity Architect, goes on to explain:  

By definition, a metasystem must be inclusive of the other underlying systems.  So for those new to the discussion, InfoCards are not positioned against any of the systems Craig mentions.  In theory you could have an InfoCard that represented an identity provider based on SXIP technology, or on Liberty technology or whatever else.  In fact a number of people are thinking about building this type of offering. [This would require adding]  a bit of code. But ubiquity and inclusiveness make such a potent combination that it would be well worthwhile.

A lot of people are mistrustful of Microsoft's initiatives in this area (and probably with good reason), but Kim has done as good a job as I can imagine being open and inclusive.  He's got the additional problem of trying to herd that management cats at Microsoft too and he's been masterful at that.  We've got a ways to go before Identity 2.0 is a reality, but I think we're making progress.

If you're interested in Internet Identity a bunch of us are getting together in Berkeley at the end of October to discuss architecture, principles, and governance at the first Internet Identity Workshop.  We'd love to have you join us and participate in the discussion.

Topic: Browser

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

3 comments
Log in or register to join the discussion
  • I know it's a typo you'll fix, but...

    ... I like this sentence:

    At Novell, Craig Burton was one of the driving forces behind the modern notion of a network as a collection of wires rather than a collection of wires.

    This is a reminder that, however conceptual what we're doing may be, the practical aspects of making something happen are still essential.

    Sometimes, in fact, the "wires" prevent grandiose conceptions from being implemented. And that can be a good thing.
    Anton Philidor
  • Digital Beer

    As good an example of how ridiculous a centralized authentication authority can be, there's one thing that those authorities provide that I personally feel is sometimes worth the hassle: Enhanced security.

    To go along with the same example, while it would be a hassle for us all to stand in line and wait for the clerk to call the DMV to verify that we were us, it would also drastically cut down on the number of underage kids using their brother's ID or a forgery.

    Some other benefits of a centralized authentication, I think:
    - Universal Trust. A single source means that there's only one place where producers must go to protect their assets by requiring identity verification. Unfortunately, this also translates into "single point of failure", but presuming the people who set up this authority were, perhaps, not-for-profit and also competent, redundancy should be available that would make the chance of nonavailability remote at worst. Of course, just make sure they have one server on the Cogent side and one server on the Level3 side.
    - Security from forgery. Or, at least, lower liklihood of forgers being able to establish an ostensibly trusted ID-publisher used for nefarious means.

    I think, in the end, that the best solution I can come up with (and I'm admittedly under-researched on this subject) is something biometric and simple, like an open-source algorithm for checksumming fingerprints.
    kgrant
  • RE: Criag Burton cries 'ubiquity'

    (<a href="http://www.blu-rayvideocovnerter.net">Video Converter</a> // <a href="http://www.blu-rayvideoconverter.net/blu-ray-ripper.htm">Blu-ray ripper</a>)
    danieu