ITANA Conference Call
January 24, 2008
 
 **Attendees**

Jim Phelps, University of Wisconsin - Madison
Paul Hill, MIT
Tom Barton, University of Chicago
Jon Giltner, Universityof Colorado
James Hooper, University of St. Louis
Hebert Dias-Flores, University of California - Berkeley
Klara Jelinkova, Duke University
Dean Woodbeck, Internet2 (scribe)

**Agenda**

(0) Roll Call. Agenda Bash.

  1. Accept minutes of last call
  2. Upcoming Meetings at Presentations on EA:## EDUCAUSE MWRC (March - in Chicago)
    1. EDUCAUSE Annual Meeting (October - Orlando, FL)
  3. Next Face2Face - Possibly June 18, 2008. Location TBD.
  4. Organizational Information Models - Tom B
  5. Data Classification and Data Retention Policies - Klara J

Items on the shelf:

  1. Future Call - Scott Converse on modified 6Sigma for Higher Ed, UW-Madison, Exec Education Program
  2. Architecture Tool discussion (All)## UC Irvine's open source tool - Protoge
    1. Chicago's I.T. Ecosystem Tool (Tom B)
  3. Paul's piece on Standards for Arch Documents -  standards for architectural documentation (Paul H)
  4. UC-Berkeley Roadmap document (Hebert)
  5. Mellon ESB Assessment - goal? is there date on this? (Mark P)
  6. Mellon New Initiative: Framework for scholarly studies tools (Keith H)
  7. Web CMS RFPs (Jim P)

(99) Next steps, next call

** Presentation at conferences**

Jim will be presenting at the EDUCAUSE Midwest Regional Conference in March about ITANA and enterprise architecture. He is also submitting a proposal for the EDUCAUSE Annual Meeting, which will take place in October in Orlando. He is asking ITANA members to provide cases concerning how they have approached EA at their universities, with the goal of having a variety in issues and approaches.
Please consider making presentations at regional EDUCAUSE and other conferences

Jim is still working on a face-to-face meeting on June 18, with the location still to be determined.
There was a general discussion about other conferences that ITANA members could recommend. Jim is attending an architecture practitioner's conference in San Francisco next week. Paul, and possibly Hebert, are attending the Catalyst conference in June in San Francisco. Sue Sharpton sent a note to the list last week about an EA conference in Orlando in April.

**Organization Hierarchies and Information Models**

Tom Barton presented information from a recent discussion at the Common Solutions Group meeting, led by Bruce Vincent of Stanford, concerning data management and organizational hierarchies. There are a variety of use cases that support an arrangement in which a number of co-existing hierarchies are recognized and supported.

For example, a person may sponsor a collaborator from outside the institution. If the sponsoring individual moves on, the collaborator may lose access privileges, even if the collaboration is not yet complete. The collaborator's access should not be tied to an individual, but to a hierarchy. Hierarchies also allow a scoping of privileges - a person may only approve account spending from within a department, for example, and not for a college.
Several hierarchies can co-exist simultaneously within an organization; for example, a chart of accounts, an HR hierarchy and an academic hierarchy. If there are many hierarchies, should that data remain in the silo or is their value in bringing it to a central location. If that is the case, who should steward or govern the process of maintaining the organizational data in a repository? And an organization structure may be operative now, but what if it changes and there is a need to go back and look at some point in the past?

Among the Common Solution members, there are only a few trying to deal with various hierarchies and governance, including MIT, Stanford, the University of California - San Diego, and the University of Washington.
Paul Hill discussed MIT's approach, which is to operate like a data warehouse, in some ways. There is a recognition that different hierarchies exist and will have different data. The concept is to place each in a database and demonstrate the relationships between the hierarchies. The central repository can draw data from the different hierarchies.

The hierarchies may not map to one-another, but you can use the data from the appropriate hierarchies, depending on the context.
Paul mentioned that MIT uses SAP, which provides a large amount of information to those who have access. Determining ways to limit that access have been challenging. Duke is facing a similar problem. Klara Jelinkova used a program for online continuing education for nurses as an example. The business hierarchy includes company codes, based on where the nurses work. But supervisors, who must approve whether someone has completed the appropriate training, aren't administrators related to these codes, because that would give them access to too much information. The solution was to create a new hierarchy for these people. There was a similar issue at MIT concerning environmental health and safety procedures and they used the same solution.

By relating one hierarchy to another, however, you can see how people from one are associated with the other. A fund center hierarchy will show the fund numbers and who has control over those funds. A a payroll system will show who is getting paid and who is authorized to change the pay rate. An academic hierarchy will show who controls different aspects of the department. You have a model for each different view.
As further examples, a dual major would appear under two different academic hierarchies.  Some faculty will appear in many departments, depending on their teaching assignments and academic appointments.

At MIT, this data is managed the same group that runs the data warehouse, even though this is considered a separate system. But, information from the data warehouse populates this database.

Paul shared a URL that includes a proposal for a master department hierarchy:  http://web.mit.edu/repa/www/master_dept.html
Klara from Duke mentioned that the political fallout can be more troublesome than the technical issues. Rather than adjusting existing hierarchies on campus, Paul said MIT creates a new hierarchy and combines new information. This takes them away from the position of potentially telling people that they have to adjust their existing hierarchy, which can lead to unintended turf battles.

Klara said the most productive approach may be to create use cases within your own institution, then work through problems at that level. For example, dealing with situations like the continuing medical education case, plus 2-3 other cases of that scope - can help make the case later for addressing the issues at an institutional level.
Given the time constraints of the phone meeting, Jim suggested continuing this conversation online and the posting of charts and graphs on the wiki. He will ping Bruce Vincent to see if Stanford has any information that could be shared via the wiki.

**Data Classification**

Klara suggested discussing how institutions classify and manage data sources that have increasing large file sizes, thanks to media-rich services. For example, does a students' senior thesis warrant long-term archiving, or does that data go away in a year when the NetID of the student is discontinued?

A number of Duke users are running up against their five-gig quota because of the large sizes of downloaded media files. People have started creating groups and uploading group-shared files, since those do not count against the quota. Once a class ends and the group expires, how long should those files be archived? In addition, researchers have immense storage needs. How long do you keep research and experiment results?
Jon Giltner said the University of Colorado is just embarking on an effort, engaging library staff to help classify such data. Some data will need to be on hand and readily accessible, for example, and other data could be moved to a lower level, in terms of storage and retrieval.

Jim will place this topic back on the agenda in two weeks. He encouraged ITANA members with specific question to have these placed on the agenda. Klara will come up with some use-cases for the group to walk through during the next call.

**Next Call - Thursday, February 7, 2008 - 2:00 p.m. EST**

  • No labels