Signet Working Group conference call Friday, September 30, 2005
*Attendees*
Lynn McRae, Stanford U. (chair)
Andy Cohen, Stanford U.
Linh Pham, USC
Andrea Beesing, Cornell U.
Stephen Barrett, Cornell U.
Butch Labrecque, Cornell U.
Joy Veronneau, Cornell U.
Karel Sedlacek, Cornell U.
Tom Barton, U. Chicago
Blair Christensen, U. Chicago
RL "Bob" Morgan, U. Washington
Gary Brown, U. Bristol
Steve Olshansky, Internet2
Jessica Bibbee, Internet2 (scribe)
Carry Over *Action Items*
[AI] {Bob} will send .htaccess local syntax to the group via the list.
[AI] {Tom} will send a few brief Signet case studies to the group via
the list.
[AI] {Group} will develop use cases for Signet.
[AI] {Jennifer} will solicit on site feedback from UC Davis about the
UI demo/mock up.
[AI] {Minh} will develop a list of requirements for how Signet will
interface with LDAP and Grouper.
[AI] {Tom, Jennifer, and Gary} will discuss the modularity of Signet’s
UI and the internationalization of code for Grouper and Signet. There
will be a separate call for this item.
[AI] {Lynn} will write up a person and function summary to express the
relationship of privileges to roles and to determine what gets
expressed in the eduPerson entitlement space.
*Discussion*
Signet v0.6 is in the final stages and a target release date within the
next two weeks is expected. New features to be included are proxy
functionality and support for internal roles – administrator and
subsystem roles. Signet v0.6 unifies assignment display code to include
proxy and allow switching between assignments you have and assignments
you gave, and eliminates a UI bias towards people who are grantors. A
logged-in user may not actually have granted nor been given privileges
– they may only be looking at assignments given to another person.
Signet release v1.0 may only include functionality for the UI, allowing
time for the API to support backward compatible functions. The Group is
still working on defining a varied set of use cases, so as to provide
corresponding solutions.
The Subject API will have its initial release soon after the Signet 1.0
release, and will be a critical addition to Signet and Grouper
alike. In the future, it might be extended to express
relationships between subjects as a possible simple approach to
supporting both tree and group structures through a common API.
At the Fall Internet2 Member Meeting, there was discussion about which
kinds of technologies or documentation structure are available. It
might be advisable to look to the future and plan on incorporating
those I2MI technologies that could benefit Signet and Grouper.
Attendees also expressed concern regarding the scope of integration for
Signet and Grouper – how well do other applications plug into Signet?
What are the complications of integrating Signet across multiple
applications?
{Tom} mentioned that a generic loader is needed to move information
from other sources into the group registry, at least until people are
writing their own loaders. {Minh Nguyen} had discussed the possibility
of having an XML revision of a subject loader for the local subject
table in Signet. {Minh} and {Jessica} are working to identify exactly
how much configuration is necessary or desirable, versus simply having
a pre-loaded template.
There was a question about terminology - whether the subject ID, opaque
subject ID and unique ID have the same meaning – they do. Consistency
in terminology across the application will be especially useful to the
users as they map namespaces across applications. Given a unique
internal key, you would have a bag of attribute with a known set of
information, such as login ID or university number. The set of
attributes would contain enough unique information that the
authenticated user can be identified, by more than simply the login ID.
It is an individual choice under the scope of interoperability, whether
there is a campus-wide ID that is shared with other applications.
Discussion turned to a situation resulting from having a requirement to
implement a Subject API binding of a local user. There needs to be a
way to associate the login ID (Tomcat) with the LDAP entry. This led to
talk of degenerate cases where, for example, in Grouper, anyone who is
the agent may be able to view things – publicly – which is an
unauthenticated situation. Is there a need for anonymous
authentication? Currently, the emphasis is not so much who you are, but
that you are able to authenticate – through login passwords or
Kerberos, etc. Gaining access to the application is addressed
separately from giving privileges.
{Lynn} asked if there was interest in or a need for a principle
identifier – if you are sharing a set of IDs for use in multiple
systems through provisioning, you may also want to know which
identifier is the login ID. This presents a use case for mapping
several identifiers.
The next Signet call will be held on Friday, October 14, 2005 at 11am
ET.