Hi Alex,
Sounds pretty interesting. I'd love to read your thesis; do you have a
pdf version you can send us?
On 2/18/13 9:54 AM, Alex O'Ree wrote:
A few years ago, my master's thesis focused on improving UDDI. The
short short version of the thesis proposed the following changes:
- Provide a simpler web service interface that just returns endpoints
- Increase security by deprecating the security service in favor of a
SAML based STS or something similar
- Increase interop/usability by defining a rest style interface and
providing hooks for the ebxml std
How do your changes line up with JAXR?
- Advertise the location of UDDI services via a multicast
Unfortunately, due to the state of the uddi working group at oasis and
the exuberant costs of membership for participation, I feel like some
my work was in vain.
Thus, I'm writing to the jUDDI group. Is there any interest in perhaps
making some of these things a reality? I'm willing to volunteer some
time to make it happen (as well as make a more user friendly
interface)
Is it even worth it? Are there any stats on juddi deployments?
UDDI never reached its potential at the 'internet' level, but it is sort
of reborn at the company level. JBoss ships jUDDI with the SOA Platform
where it is used as the central service repository.
The trick with adding these features will be to add them in a such a way
that we can still be spec compliant; so they have to be 'on top of' the
current infrastructure or will undermine the adoption of jUDDI itself.
People like to be vendor neutral; which why there is UDDI specification
to begin with. On top of building on top of the spec if add these
features we'd have to document how other vendors can implement these
enhancements.
Looking forward to discussing this some more!
--Kurt
