October 13, 2005

Members present : John Board, Shailesh Chandrasekharan, Wayne Miller, Angel Dronsfield, Michael Holt, Nevin Fouts, Tracy Futhey, Michael Gettes, Daron Gunn, Billy Herndon, Roger Loyd, Gregory McCarthy, Kyle Johnson, George Oberlander, Mike Pickett, Rafael Rodriquez, Heather Flanagan, Joseph Meyerowitz, Molly Tamarkin, Christopher Timmins, George Turner, Robert Wolpert, Steve Woody

Guests: Tim Poe, OIT; Allan Blatecky, RENCI; Tammy Closs, OIT; Ginny Cake, OIT; Jonathan Adams, OIT

Start time : 4:03 p.m.

 

I. Review of Minutes and Announcements:

  • Explanation of phone problems with Athletics: Tammy Closs says Athletics was overwhelmed with calls about basketball tickets, which resulted in trunk problems. The issue has been resolved, and we're working on better communication with Athletics to help them understand that this is not the first time this has happened.
  •  

II. Introduction to Tim Poe, Sr. Manager of Video & Audio Services: Early observations about Duke and trends in video & audio services - Tim Poe

Tim Poe says the services we are planning on providing are video and audio streaming, video on demand, and video conferencing over IP and all of the infrastructure involved there. We want to have things in place to have multiple people involved, so you can look into a directory, see their e-mail and phone numbers, and have a gateway in place for telephony and video conferencing. We're also working on video conferencing from a variety of venues, lecture capture, and consulting.

Web portal: UNC has a knowledge management tool, help.unc.edu, with tools for easy entry, XML to pull out and put in your own school wrapper. It makes it easy to organize content for help and organize training. What is the central clearinghouse for training? It doesn't matter where it comes from; is there some sort of consolidated tool that manages that? That is something else UNC does well.

Support of broadcast-quality video: To get to CNN, etc., we need high quality. What are some of the transport mechanisms we can make available?

Tracy Futhey says this becomes more pressing as the university and Office News and Communications take a bigger role to hook-up leaders with news media.

Tim Poe says for outreach, support, and documentation synchronous learning tools like Breeze have access to tools in a virtual meeting. Is there going to be an enterprise solution for that? The lines get blurred with VoIP. We're looking at subjective monitoring. For example, the Duke Chapel has a live feed during services; how do we monitor them, let people know it's going out effectively, and if not figure-out how to do something about that?

Robert Wolpert says I would see some things that I wouldn't see as necessarily helpful; there is going to come some need to beat the bushes, I see a balancing act.

Tim Poe says we will be seeing which groups have listserves and looking for guidance. We have high expectations and high needs for the group. We are open to suggestions.

Robert Wolpert says I encourage you to reach out to the faculty.

John Board says Lynne O'Brien and CIT people are a conduit for innovative proposals for how to use video.

Tim Poe says I have been interacting with Amy Campbell of CIT. Are there venues to go directly to the faculty?

Robert Wolpert says ITAC is a good start, same aspects for Julian Lombardi.

Tim Poe says VoIP needs to be a close coordination, but it is not under this group. Video production also is not part of this group's scope. We will be working with Johnny Bell and Kevin Miller on video furnace.

Tracy Futhey says he will be working with people to find out who are the right people Tim needs to meet and what we need to be doing.

III. The Renaissance Computing Institute (RENCI) and its relationship to Duke - Allan Blatecky, Deputy Director, RENCI Discussion

Allan says RENCI is basically funded by Duke, NC State, and UNC-Chapel Hill. Our oversight board basically consists of just six people at this point, two from each school. We're fairly new, almost a year and a half old. Dan Reid started RENCI a year ago, I came six months later, and we started hiring folks in the last year. We have a small staff of 10-12 people, students and some grad students. A lot of the projects Dan brought with him. Our basic focus is to try to raise the level of research capabilities in the Triangle. How do we raise the level of research to make this area more of a national player? One key idea is everything we do is multidisciplinary. The second component is everything is multi-institutional. We don't want to help a particular researcher get a grant.

At this point Chapel Hill is putting in more than other institutions, and we're based there now. We are trying to get a core facility in Chapel Hill . We're moving into a new facility in January with 50,000 square feet, and that should be pretty full by next summer the way we're going. We're going to be coming to campuses more and more to look for ideas. As Dan says, we want to be a place where if someone has a crazy idea and we can say, ‘Hey, that sounds like fun.' One of the things that is clear is that if you look at science, almost all science relies on collaboration and it all relies on data - how do we manage it? We're trying to help that side of the equation.

To give some examples, we have a proposal to create a collaborative facility at Duke, State and Chapel Hill . We're interested in the whole range, humanities, sciences, etc. One of the clear winners in North Carolina is the biology side between Chapel Hill and Duke and their medical facilities.

We got funding from the state to create a statewide presence for RENCI. The focus is to take research and drive it into economic opportunities. What is key is a couple things: we have proposed to cross a number of disciplines as well as areas of the state. We've talked about hurricane prediction and surge modeling. The idea is to apply surge models to North Carolina so we can look at disaster planning. We can look at multiple projects and issues like this to look at statewide planning.

John Board asks can you add any exemplar problems being worked on?

Allan says if you have 10,000 computers and one is always down, how do you manage that? Also, there are a couple of projects in genetics with some of the NIH activities about how to manage big databases. NIH, especially on the medical side, has lots of data that is not shared. The issue there becomes how to begin to address these issues. We're doing a lot of work in portals and grids because we're getting to a point where we can do nice things with frames.

John Board asks what is the actual model for collaboration.

We act as the coordinator. For example, if Duke is the lead, we bring the IT component in and do outreach. It's not really administrative. Right now we are limited with staff, we don't have enough people, so when get more folks we want to engage the faculty more.

Robert Wolpert asks how would a researcher discover they need to work with you?

Allan says if we get presence on campus it will greatly help. Now we are in Chapel Hill and it is still a problem. We're trying to get some visibility.

John Board says a year from now, if a faculty member with an idea that has some IT component comes to you, how will that work?

Allan says hopefully we will have someone on campus that can make those connections. The idea is to make things bigger, to look at some larger, longer term grants.

Molly Tamarkin says it sounds like a facility that faculty at the Nicholas school could use, you think of mapping, etc., but I suspect it will be hard for some faculty to take advantage because your computing infrastructure is so different from ours. Migrating to a different level of computing would be the catch for us.

Allan says the point is, how do we do that, what is the science? That's the driving issue.

John Board says since you have an interest in economic development, some could see it as pressure to create spin-off businesses.

Allan says we're not looking at new companies, though if that happens that's fine. Economic development at the local level isn't infrastructure, it's providing them with experience. It's tougher, but fits in with the educational mission better.

John Board says at one point it was thought some of the RENCI staff would be floating faculty with special knowledge in, say, data storage.

Allan says we have some space for visiting faculty. The problem is telling people if they come to us, they have to collaborate.

Robert Wolpert asks is there any collaboration you can see with NIHS, EPAs, or with corporate research centers in RTP?

Allan says of course, we want to see how to get NIHS. If we can get NIHS hooked up we may get NIH involved. The problem we're facing now is we need space on campus.

IV. Telcom Renovation Update - Michael Gettes

Michael Gettes says the Telcom renovation update was institutionally initiated as of May 2005. We performed a feasibility study and had to get it all done by the September board meeting. We presented to Dr. Trask, Tracy, and John Pearce in August. We’re going against the Uptime Institute, looking at data centers from this industry standard, which is a four tier perspective. We’re looking at the Telcom Building as a Tier IV minus.

Highlights: Currently the DHTS space is between Tier II and III; the OIT space is between Tier I and II. There is approximately 18,000 square feet for computing of a total 26,500 square feet, and we will use the North Building for redundancy until we figure out the future distribution. There is space for the network and for departments, supply racks and cooling so people can just bring in machines, even lockable rocks. The traffic flow is for system administrators so they can only gain access to that space and not to other restricted spaces in that building. The first floor is to be one large space. The rest is for DHTS.

Human traffic flow: The spaces are properly controlled, the building is properly controlled, the public space is only public if you can get into the building, and after that it’s all controlled. There is no “front entrance” for the building. The only entrance is through the loading dock or ends of building where the stairways are.

Floor to ceiling clearances: The ground floor is 10’4”; we need to depress the floor by 24” to install the 24” raised floor. The first floor is 11’2”, which will leave 26” above racks. The second floor has a 13’ clearance, which will leave 4’ above the racks.

Elevator: The current elevator has a load of 3,500 lbs, so we will install a custom elevator to fit the existing shaft to have a 4,100 lb load.

Generator: The existing generator is 30 yrs old and is at the end of its life. The capacity is to be increased to 1200kVA to meet the backup needs of entire building with a 40-hour run-time at full load. That might increase, because there might be some issue of understanding the true size of the tank underground.

Electric, UPS: There is a pair of 500kVA UPS on each floor, and a few housed in the panel board. 420 tons of capacity are needed for HVAC. The primary cooling is chilled water; the secondary system is air cooled with units located on the roof.

Structural: Tiers III and IV stipulate a minimum load of 150 lbs/sf. This study doesn’t reflect a change to support 300 lbs/sf on half of the first floor for HPC. The ground and second floors will be at 150 lbs/sf.

People impact: The telephone operators and the networking and telcom store rooms are still there. We’re they’re going, we’re still investigating options. The total project cost is estimated at $8.6 million. That includes 30% contingency and costs for temporary power and cooling associated with existing operations. Funding will come from central, OIT and DHTS.

Nevin Fouts asks what is the timeline to complete the renovation?

Michael says I was told in May that it would take nine months to get where we are now, and it took about three months. The real challenge is based on how things are normally done and how architects normally work. We’re not at the drawing phase, and we need to get to that point.

Shailesh Chandrasekharan asks has it been decided how to allocate spaces for departments?

Michael says no. That’s going to involve participation of things like ITAC. We just wanted to make sure we could offer some amount of space where people could go and have good computing space.

Tracy says with Pratt we’ve been fortunate to have the opportunity for near term access to get us through the next few months. CIEMAS came on a year ago and hadn’t fully populated its machine room. Pat Driver is working with Pratt and DHTS to figure out how to fully equip the space so many of us could occupy it until we move into the new space.

Michael says also, for high performance computing we have space allocated in the North Building, which was renovated this summer, so we have some expansion room there.

Tracy says Michael’s urgency to complete the renovation isn’t that we don’t have place to put things today, but that 17-18 months from now, if we don’t do the work we need to get it open, we’ll really be in bad shape.

V. Group Email Policy Update - Tracy Futhey, Heather Flanagan

Tracy says we had a couple of mailings that went out that were problematic because of the way in which they happened. One was related to an e-mail about the Rolling Stones concert – it was sent in html format and many people who received it on campus couldn’t read it. Heather is distributing a proposed modification for that and a couple other policies.

Heather Flanagan says the last time the group e-mail policy was modified was when it was first published back in 2001. What doing now is trying to clean up what left out initially, explaining what the policy is intended for and when a message is of common importance. We’re trying to see e-mail as a fairly sensitive media for sending out information.

Tracy says the first modification tries to tell people not to send random e-mail, with clarifications about who approves things. There is a specific one to address the notion of plain text as the only thing we can really trust, and further reference that even if things are initiated within departments, there are some things it is good for us to know about. This is the e-mail equivalent of the problems with the phone system when Athletics opened the floodgates for tickets.

Heather says one thing we don’t do is get approval for people submitting the request.

Robert Wolpert asks is it really better to spam people using paper mail?

Heather says we can’t be a clearing house for how people communicate; we don’t control the paper mailings, but we do have a say in the e-mail policy.

Nevin Fouts says there are a number of principles in terms of approval. In our school most communications are in the many hundreds because there are so many people in the school. Do the limits apply?

Kyle Johnson asks what is the expectation for how far in advance someone should give you notice?

Heather says if you ask OIT to do the mailing, give two business days lead time. This is sometimes not possible, like for the Bursar. If you are doing it yourself that day, the notification is more of a heads up.

Nevin says there is a message every day from some part of the administration to some part of the school, so almost every message would exceed threshold.

Kyle says in Student Affairs Dr. Moneta has approved a class of messages, so if we have operational messages going out, we don’t email before each one goes out.

Heather says I’d be a little worried if daily a 1,000-recepient message was going out to specific people with attachments.

Molly Tamarkin asks do you think you need to say something about not being responsible for the content of the mail?

John Board asks do we need to approve this?

Tracy says would like to have ITAC say it’s approved.

VI. Common solutions group meeting update - Michael Pickett

Mike Pickett says Tracy, Michael Gettes and I went to Carnegie Mellon for the Common Solutions Group (CSG) meeting. CSG has a long day workshop, then a short half-day workshop, then to a meeting and policy discussions.

Twenty-five universities participate: Duke, UVA, Berkeley, MIT, Harvard, lots of the technology powerhouses. Invitees include the CIOs and their chief technical types. It is intended to find common solutions, though we typically spend a lot of time talking about common problems. All the presentation information can be found at stonesoup.org.

The long part of the meeting was on a Fedora repository, out of UVA with a Mellon Grant. The idea is to have a huge repository of all types of teaching objects, materials, paintings, etc., and to limit access to students over particular time giving lots of control over copyrighted data.

There was a survey of CSG member institutions by Tom Dopirack at Carnegie Mellon. There are ten schools with a CMS, and most are commercial. His thought: it’s not all perfect yet.

Digital libraries: Michigan did a discussion on what they are doing with Google libraries. Google Print is beta application that allows you to put a phrase in and allows you to see all books that have a line with that phrase. There is some controversy about if this is good or not. Google is doing this, and will give the entire content back to the University of Michigan.

I was one of the co-runners of the short workshop on enterprise data. There was a panel on information security disasters in higher education. We talked a lot about prevention, identity management, and a how directory services can be a preventative service and also how we have to be careful with that info.

Information in terms of trending: There are more attacks being done for money than not - people are selling access to machines for launching attacks or sharing info. There’s not a lot of evidence yet that people are mining data on machines that had been broken into.

The big trend now is from the federal government that will require large research projects to a have plan for how protect data and how to share data.

Should we get rid of student e-mail? Most institutions said no, a couple said they are looking at it very seriously. Most people said we depend on student e-mail to be absolutely sure we are communicating with students. We can’t outsource that.

VI. Other Business

CIEMAS now has cell phone coverage inside.

End time : .5:32 p.m.