*Minutes*
*Salsa-DR Working Group *
*May 23, 2008*

***Attending***

Don MacLeod, University of California-Berkeley (chair)
Rodney Petersen, EDUCAUSE
Carol Templeton, University of Tennessee
Bill Owens, NYSERNET
Walter Patruska, University of San Francisco
Joe St. Sauver, University of Oregon/Internet2
Steve Olshansky, Internet2
Dean Woodbeck, Internet2 (scribe)

***Action Items***

[AI] (Don MacLeod) will provide a report on UC-Berkeley’s DR test, which is scheduled for the first week in June. He will provide the report during the next call.

[AI] (Carol Templeton) will provide a report on the University of Tennessee’s DR test, which is scheduled for the last week in June. She will report during a future call.

[AI] (The group) is to think about and discuss (via the email list) how a nationwide DR preparedness event might be organized for September and whether a scaled-back version might be more manageable.

***Carryover Action Items***

[AI] (Don MacLeod) will develop a list of issues for people to consider when considering outsourcing hosted email.

***Spring Member Meeting Update***

The Disaster Recovery BoF at the 2008 Internet2 Spring Member Meeting had a good turn-out, with plenty of good discussion. One proposal that resonated was a nationwide testing/evaluation of disaster recovery processes. The thought is that a consolidated day would be more realistic, in the event of a large regional disaster. Companies that serve as out-source solutions, for example, would be tested to see if they could withstand the demand when several of their clients require services all at the same time. The thought at the BoF was to schedule this for September.

***Disaster Recovery/Business Continuity Tests Planned***

Don reported that the University of California-Berkeley will perform an annual DR test during the first week in June. He characterized the test as more of a system recovery test for select systems. Berkeley’s off-site DR vendor is IBM in Boulder, Colorado. [AI] (Don) will report on the results during the next DR call.

The University of Tennessee, which also uses IBM’s facility, will conduct its test during the last week of June. Staff members from Tennessee will travel to the Boulder site to participate. [AI] (Carol Templeton) will provide a report on the test. She will report during a future call. The University of Tennessee system is looking at ways that the five institutions could work together to provide back-up services and redundancies.

NYSERNET has developed a centrally located data center for its member-institutions, located in Syracuse, New York. The center was designed to satisfy institutional disaster recovery needs and has the capability to provide such systems as enabling critical communications in the event of emergency, as well as the ability to duplicate multiple institutional administrative computing systems.

***Coordinated Disaster Recovery Testing***

The group discussed whether there was momentum for developing a nationwide disaster recovery preparedness event. Because Internet2 and EDUCAUSE would have limited resources for supporting such an effort, any such event would need to be led by someone from the community. There was some concern among group members over trying to develop such a program in the short time between now and September. [AI] Don asked that the working group think and discuss (via the email list) how such an event might be organized and whether a scaled-back version might be more manageable.

***Partner Institutions in Disaster Recovery***

The working group discussed issues that emerge when working with partner institutions for disaster recovery. In such cases, institutions agree to serve as back-up resources, sometimes hosting hardware that is owned by the other institution. There was a discussion about how priorities would be set if both institutions had failures at the same time. Many institutions operate on a handshake, as opposed to a written agreement, so such situations can become fluid.

There was also a discussion about not having all of your disaster recovery eggs in one basket. Some institutions have different back-up sites for different systems. The administrative system, for example, may be backed-up in a different location than the student system or the email system.

***Virtualization in Business Continuity***

Carol reported on a concept emerging at the University of Tennessee that the five campuses serve as virtual DR/BC centers for one-another. There is concern, for example, that the financial system for all of the campuses is in Knoxville, using 30 servers. There is some interest in having a virtual back-up of the system at one of the other campuses. Don reported that some University of California campuses have also begun thinking about a similar arrangement, and about building separate data centers to serve all of the campuses.

The University of Oregon is working at replicating two sites on campus, for redundancy in the event of hard drive failures.

Don proposed inviting a speaker to a future phone call, from a campus or from a vendor, to provide insight on the types of DR/BC facilities institutions are using. Bill mentioned that DR and BC are the topics for NYSERNET’s next community technical meeting on July 29.

***The next call is June 20, 2008, 1 p.m. (EDT)*