UC San Diego is hosting CSG for the first time.
Ken Klingenstein (Internet2)
Agenda for the morning:
Use Cases: LIGO, iPlant, and Bamboo
Basic generic build issues/COmanage
Basic outsource issues/Google offerings
The social networking angle
Connect/Sharepoint experiences
Panel
Federated Identity – reducing the authentication barriers to collaboration, note that sometimes the IdP is not the enterprise but the VO or other org.
Multiple levels of assurance, depending on use.
A new word: Domestication of Applications – about refactoring applications to use the emergent identity services infrastructure. Make an external call.
Begins with federated identity and authentication, but gains a lot from group management for access control, etc.
Lots of different flavors.
Examples:
COmanage – the Dutch have done amazing things
Commercial offerings – Sharepoint, Adobe Connect, Google Sites, Wave, Google Apps
Repurposed LMS – Sakai, Croquet
Dutch National Collaboration Infrastructure – serving a million users
Domesticated tools – Adobe Connect; Alfresco; Foodle; Filesender; Confluence, Drupal, etc. Google Apps; My Experiement.org.
Done both grid integration and workflow.
Key issues:
Extent of application domestication
Appliance, service, cloud offering
Waiting for other technologies to happen – interfederation, discovery, metadata tagging, etc.
We’re early in understanding the UI
Domain applications/ science portal – can I use my groups for getting to grid application?
Collaborations and Virtual Organizations
– Move from a tool-based identity world to a collaboration-centric space
Roles, schema, and attributes
Big Science Collaborations
LIGO (ligo.org)
The single largest VO funded by NSF – does high energy physics.
Complex internal access issues – lots of internal competitive proposals requiring complex access control
A small number of very large files
iPlan – http://www.iplantcollaborative.org
NSF leading with cyberinfrastructure for the plant biology community
Broad outreach, education and training components – rich external access issues.
A very large number of small files
Needs of Big Science Researchers
Access to collaboration tools
No modifications to existing domain science apps – in some cases jobs run for years.
Command line tools – an interesting challenge
Internation capabilities
Multiple levels of assurance
Roles, attributes, metadata, and ontologies
Chad Kainz (U Chicago) – Humanities Research – A View from the Bamboo
Google is a big humanities project š
Tow facets of collaboration – substantive and methodological – Methodological is a common goal. Substantive is uncommon – “I’ve got my thing and I want to unleash it on the world”. Strong desire to shift towards methodological.
Scholarly Networking is not the same as Social Networking
Social networking is focused on individual trying to connect with known group (friends, colleagues, family)
Scholarly Networking – Individual is seeking different connections that cross disciplines or engage other individuals with similar interests elsewhere.
What is emerging are invisible colleges of like-minded individuals who work at different institutions.
Pub problem – Neet to be in the right pub at the right time to make the right connection. “The most expensive dating service we have on campus is the VP of Research”
Five things came out of Bamboo workshops:
– Enable the discovery of scholars and their work at scholar-scholar level. Requires contextual metadata about projects, content, services, etc. Manual entry of metadata will fail (duh).
– Enable the creation of scholar profiles from data sources at institutions and create mechanisms to mine the data across institutions. Creates tension among institutions, scholars, and scholarly societies (academic “stars”).
– Organize – enable groups to organize outside of normal boundaries.
– Engage – enable a variety of scholars and institutions to engage in the network even if they don’t have organized data.
– Market – create a participatory market to promote greater SME interaction.
Ken – In LHC crowd, “discovery” is finding 2,000 processors they can use. In humanities it’s “is there a piece of software out there anywhere that can do…?” Distinction of provenance – in big science it’s store data so experiments can be repeated in the future. In the humanities it’s about giving proper credit – building arguments on arguments, so need to be able to work your way back through the subjective views. If an assertion is made by a grad student, who’s the advisor of that grad student?
A Middleware Unified Field Theory – Mike Gettes (MIT)
This is about Internet2’s COmanage.
We want inter-enterprise workgroup collaborations (or CO – Collaborative Organizations)
Identity, groups, federation, and applications.
Give control to community members – stop making people come to central IT.
Integrate with existing higher ed infrastructure.
Shib is federating technology. Group management. LDAP-PC publishes to LDAP, apps talk to LDAP.
Foodle – A federated Doodle.
Google Wave – Chris Hubing (Penn State)
Use cases: collaborative authoring of student documetns – teacher can play back wave and watch evolution of document; wavelets for discussion generation – hard to in google docs
Federation steup “like email” dns srv records; uses x.509 certs between wave servers; wave fed protocol is an extension of XMPP protocol.
Wave Fed prototype server – plugs into XEP-0114 Extension.
Wave providers right now or FedOne and Ruby on Sails (written by a high school student!)
Federation only works with wavesandbox.com – main google wave site doesn’t do federation yet
only trusts StartSSL certs – a little questionable on business practices
no web ui for fedone prototype server
if you use google apps for edu you need to disable chat service, because of name service collisions.
Oren Sreebny 