On Sun, Nov 16, 2008 at 2:33 PM, Arnaud Lesueur
<[EMAIL PROTECTED]> wrote:
> Scott,
>
> I share your point of view for those modules migration. And I will be glad
> to migrate the SPNEGO module as long as you do not push me dead line :-)

We've got plenty of time... API's haven't been finalized yet either ;-)
>
> Few remarks :
> - Trusted support is really important for intregration of CAS with others
> products and should be kept

I agree.  I'll be making sure that gets transferred over.

> - Handlers should be able to deal with password expiration natively if it is
> possible (specially the LDAP handler)
Yes, I agree. I also want a way for us to send back information like
"you'll password will expire in 10 days."

> - Radius Handler is really usefull for integration of CAS with OTP systems
> and I would also like to see it in CAS 4. I have use it yet but I will try
> to have a look at it for improvrements

Feel free too...we don't use RADIUS as our authentication mechanism here...

> - Clustering support is a bit tricky. I know that you had better results
> with memcached. But, this suppose to deploy/maintain/monitor an extra
> non-java components on cluster nodes

Correct.  However, I only know about two places that had good luck
with the JBoss Cache :-)  At a minimum we should upgrade to JBoss
Cache 2.0 and take advantage of their more fine-grained controls.

-Scott


>
> Regards,
>
> -Arnaud
>
>
> On Thu, Nov 13, 2008 at 4:56 PM, Scott Battaglia <[EMAIL PROTECTED]>
> wrote:
>>
>> I should also add that just because Rutgers is interested in migrating
>> certain modules, that does not preclude anyone else from expressing
>> interest in maturing/evolving those modules also (or in place of
>> Rutgers).  I know there are some people that have great ideas on how
>> to improve the LDAP handlers.
>>
>> -Scott
>>
>> -Scott Battaglia
>> PGP Public Key Id: 0x383733AA
>> LinkedIn: http://www.linkedin.com/in/scottbattaglia
>>
>>
>>
>> On Thu, Nov 13, 2008 at 10:52 AM, Scott Battaglia
>> <[EMAIL PROTECTED]> wrote:
>> > All,
>> >
>> > As part of the CAS4 roadmap/development, I'm looking at the currently
>> > available modules/add-ons for CAS3 and evaluating which should be
>> > migrated to CAS4.
>> >
>> > We currently have the following modules (not including core and webapp):
>> > * Compatibility
>> > * BerkeleyDb Support
>> > * JBoss Cache Support
>> > * Memcached Support
>> > * RESTful API via Restlet
>> > * Generic Handlers
>> > * JDBC Handlers
>> > * LDAP Handlers
>> > * Legacy Handlers
>> > * OpenId 1.1 Basic Support
>> > * RADIUS Support
>> > * SPNEGO/NTLM Support
>> > * Trusted Support
>> > * X.509 Support
>> >
>> > Currently Arnaud handles the SPNEGO/NTLM support and I'd be extremely
>> > happy if he migrated that to CAS4. Similarly, Velpi currently manages
>> > X.509 and I'd like to see that migrated.  Rutgers specifically has an
>> > interest in ensuring that Memcached support, RESTful API, and LDAP are
>> > migrated.  In addition, I would be willing to ensure that the
>> > "Trusted" support would be migrated as its a key integration point.
>> > (in addition, not listed here, but included as part of core would be
>> > the JDBC Ticket Registry)
>> >
>> > That leaves the following modules:
>> > * Compatibility
>> > * BerkeleyDb Support
>> > * JBoss Cache Support
>> > * Generic Handlers
>> > * Legacy Handlers
>> > * JDBC Handlers
>> > * OpenId 1.1 Support
>> > * RADIUS Support
>> >
>> > Intuitively, it would seem the "Legacy" support for the CAS2 Password
>> > Handlers has minimal usefulness at this point (CAS3 has been out for
>> > over 3 years).  Similarly, the generic handlers are essentially
>> > enhanced test handlers.  Unless someone had a strong interest in
>> > supporting them, my inclination would be not to migrate them.
>> >
>> > The RADIUS and JDBC handlers seem like they would be useful to
>> > continue supporting.  Is anyone interested in maintaining them (i.e.
>> > taking ownership of them)?  They'd most likely be migrated either way,
>> > but merely as is instead of with improvements that may be
>> > necessary/useful.
>> >
>> > With the understanding that JBoss Cache underperforms compared to the
>> > Memcached client, is it worth transferring JBossCache over?
>> > Similarly, with the JpaTicketRegistry, is it worth having BerkeleyDb
>> > also? (it may since BerkeleyDb is long term storage on disk).
>> > Clearly, if anyone steps up, even if I don't find them useful, we'd be
>> > willing to migrate.
>> >
>> > Thoughts? Comments?
>> > -Scott
>> >
>> > -Scott Battaglia
>> > PGP Public Key Id: 0x383733AA
>> > LinkedIn: http://www.linkedin.com/in/scottbattaglia
>> >
>> _______________________________________________
>> cas-dev mailing list
>> [email protected]
>> http://tp.its.yale.edu/mailman/listinfo/cas-dev
>
>
>
> --
> Arnaud Lesueur
>
> LinkedIn: http://www.linkedin.com/in/lesueur
>
> _______________________________________________
> cas-dev mailing list
> [email protected]
> http://tp.its.yale.edu/mailman/listinfo/cas-dev
>
>
_______________________________________________
cas-dev mailing list
[email protected]
http://tp.its.yale.edu/mailman/listinfo/cas-dev

Reply via email to