OASIS eXtensible Access Control Markup Language (XACML) TC

 View Only
  • 1.  complexity going forward

    Posted 04-22-2003 18:43
    
    I've been spending the last several weeks working
    with XACML from the point of view of a policy writer
    and a systems programmer (as opposed to my previous
    point of view as implementor), and I thought I'd
    share two observations.
    
    1. XACML is very cool. There are a lot of powerful
    features in this language that really work. I was
    able to build a lot of complex policies without
    extending too many pieces...just custom code to find
    policies and attributes, which is needed in almost
    any system. So far everyone who's seen the demo has
    been way impressed by what XACML can do.
    
    2. Working with XACML is really hard. The tradeoff
    to building flexible, expressive, generic languages
    is that they can be hard to work with, and XACML is
    definately that. It becomes even more of a challenge
    when you start to build up real systems around XACML
    and you have to figure out how all the pieces fit
    together and how one change can affect many
    different policies and queries. I bring this up
    because I'm seeing a lot of new work items in the
    TC, and some of them look like they'll make things
    even more complex. Don't get me wrong, I'm all for
    the standard pushing forward and the right features
    getting added to the system, but there isn't a lot
    to help people along, and so I wonder whether users
    need a little bit of support before the standard
    should push in too many more confusing features.
    
    I'm particularly curious about demand for some of
    the features being discussed. I'm seeing a lot of
    use cases, but none of them seem to be backed up by
    real-world scenarios or comments from people
    actually using XACML in anything. So far my
    experience is that the language has almost
    everything that I need, and that seems to be the
    common opinion that I've heard from others too.
    Again, I wouldn't want to slow down the progress
    being made here, and I'm not trying to question any
    particular feature, but I'm wondering how many new
    features are coming from "it would be cool if we
    could do this" versus features that are coming from
    real demand. Given how complex XACML 1.0 already is,
    I would hope that most features that get added
    quickly fall into the second catgory.
    
    Anyway, I just wanted to bring these issues up since
    I haven't really seen them discussed much on the
    list. Apologies if I've made anyone upset...I really
    just want to understand where a number of these
    features are coming from.
    
    thanks
    
    
    seth