Duke ITAC - January 25, 2007 Minutes
Duke ITAC - January 25, 2007 Minutes
January 25, 2007
Owen Astrachan, Pakis Bessias, John Board, Tammy Closs, Ken Hirsh for Dick Danner, Brian Eder, Nevin Fouts, Tracy Futhey, Christopher Gelpi, Susan Gerbeth-Jones, Michael Goodman, Billy Herndon, David Jamieson-Drake, Ben Riseling for David Jarmul, Julian Lombardi, Roger Loyd, Tim Bounds for Caroline Nisbet, Lynne O’Brien, Mark Phillips, Mike Pickett, Rafael Rodriguez, Molly Tamarkin, Tom Wall, Robert Wolpert
Guests:Jim Roberts and Julia Trimmer, provost’s office; Kevin Witte, Debbie DeYulia, Klara Jelinkova, Chris Cramer and Ginny Cake, OI
Start time : 4:05
I. Review of Minutes and Announcements:
Klara Jelinkova – Heather Flanagan has moved on to Stanford. We’re seeking a new manager of Collaborative Systems.
John Board – Tallman Trask will be at our next meeting. If you have any specific items, it’s better to give us advance notice. Peter will be here March 8.
Vista update – Debbie DeYulia – What OIT is doing in technical support for Vista – We’re not officially ready to make a statement. That will be forthcoming. But a few things: We are actively testing the site license software, and administrative and enterprise software.
The one point I’ll make is the TAP computers will come with Vista. The rationale is that we know the students are going to come with that and if we don’t offer that they won’t buy through us and we’re trying to streamline support costs by using this program. That was the decision that we made as a team. They can order those as early as May 1.
The labs won’t be upgraded until at least the first service pack or summer 2008. Some departments that OIT supports, they will not be updating any time in the near future. We will have a statement coming up in the next month or so.
Mike Pickett – Was there any software not working?
Debbie – VPNs aren’t working and some anti-virus software. We have a small team looking at antivirus anyway because it’s up for renewal.
Chris Cramer – McAfee’s new version will support Vista. Cisco will have to upgrade their VPN. There are a few concerns I have, which is why I would personally hold off until July or August. The security issues haven’t been ironed out. And there turned out to be some scary issues with digital rights management.
Question – Why wait until summer 2008?
Debbie – I think the issues to be worked out by then.
Question – Do you anticipate any interoperability issues between XP and Vista.
Debbie – The tech team is looking at that, can they go backward, etc. It’s being researched.
Robert Wolpert – Are you looking at Office at all, the new version?
Debbie – That is in testing. We’re on this with other universities that are researching it.
Debbie – And Mark Minasi will be here next week. We’re going to send something out on the university side. He’ll be doing Vista training; he’s coming to do a two-day seminar, not hands-on, but training at American Tobacco. We’re hoping to videocast it. We may be doing it at the North Building. We’re checking with the health system to see if there’s a place there where we can show it, for people who don’t get a place at ATC.
- John Board – We’ll be having a significant but brief system outage to deal with some Cisco vulnerability issues
Chris Cramer – It’s a denial of service vulnerability. If we don’t reboot it, someone else will.
- Julian Lombardi – Google has just announced an award for access to depersonalized data. They’re asking for two-page executive summary, then they will invite people to provide more complete proposals. You can ask for more info at email@example.com. Bob Price is running that.
- David Jamieson-Drake – Bob Newlin had back surgery and it went well. He’s home recovering.
II. Faculty Systems update and introduction to Julia Trimmer – Jim Roberts, Julia Trimmer, Billy Herndon
Jim Roberts – I’m circulating two documents that describe where we are. The basic rationale is that over last five to six years we have put in place robust systems for many institutional processes – HR, finance, etc. – but tools for faculty systems have been done in silos. So we have an abundance of faculty systems and none of them are related to each other. We want to elevate this to an enterprise approach, raise the visibility of the project.
The provost has presented the need to the deans’ cabinet and we have good buy-in to find the best about each existing system and put them to use for us all. We are further along at defining a need than we are in defining a structure. Julia will talk in a minute about what she’s doing to get her feet on the ground.
A steering committee has been formed for this project. The provost has said he needs to own this, so he’s asked me to be the functional spokesperson. He’s created a steering committee that’s a robust group representing the cross-institutional nature of this endeavor.
The reason I passed out Julia’s job description is to give a sense of the granularity of the job. We see this role as an advocate for the sensible application of technology for administrators’ ability to support faculty. We don’t see it as strictly a come in, manage the project and go away position. She’s an advocate for good ideas to make the work of those of us who support the faculty more effective.
So I ’ll introduce Julia, who’s been on the job almost three weeks. She comes to us with a lot of Duke experience in ADG systems in the School of Medicine. Their mission has been to support academic processes, so she’s familiar with the work of faculty, research, and all the components that go with us. She’s a talented and creative project manager.
Julia – I’m happy to be here. I’m looking forward to working with a lot of these solid and creative systems that are being used on this side of the university, too. I’m looking forward to working on improving systems. I worked on the faculty side of ADG systems, Faculty Central, the faculty research directory. I’ve also worked on STAR, space accounting system, and the medical center’s IRB system.
I’ve been talking to people about the faculty database system and other systems. I’m looking for more input. I haven’t talked to any faculty yet. We’ll be setting up more formal communication channels, possibly through ITAC. We will have a project website soon that will have updates. If anyone wants to call or email me about what we’re doing, I’ll be happy to talk to you.
Jim – It’s an evolutionary process, not a big bang revolution. We can look at it as concentric circles or building blocks. Rooting the core data in the systems of record, then to appointment histories and things we need to keep the board informed, making research known. We are still fact-finding to get the building blocks lined up.
Robert Wolpert – I encourage you to speak with faculty and I’ll acquaint you with some minefields you’re about to find.
The faculty database systems have two aspects. The bright side is that you would like to keep consistent information in one place. There is a fear, however, on the faculty part that the underlying reason is to make it easier to make assessments of faculty in ways that are naïve, such as counting things. The fear is that the subtleties of people’s work will be ignored and decisions will be made without that information.
A related feature is that this is supposed to be a way of simplifying info, but it turns out that these ungrateful faculty people have their own ideas about form and style of these documents (CVs, grant proposals, etc.) and like to tailor them for different purposes. All of that subtlety is lost in the mechanical system, even if it’s possible to make some changes. So it will end up being a shadow system and a tax on their time because they’ll be keeping their own documents on the side anyway.
John Board – One of systems we’re using was a CMS for websites. Are you looking for one tool that’s a CMS and a management tool for the university?
Jim – It wouldn’t necessarily follow that it will be in one package.
John – We have three homebrew systems. Are we tracking commercial and open source developments as part of this process?
Billy Herndon – We haven’t drawn any conclusions. Some schools have done some development. We’re not at that point yet, so everything’s open.
Chris Cramer – The system will be feeding from systems of record. Will this be a system of record itself, or will it be self-contained?
Jim – We have to be agnostic about that as well.
Billy – The agreement we have is that we want to have core data maintained in the SAP system. We don’t know if it’ll make sense to add some of that. We want to make sure it’s integrated.
Molly Tamarkin – If you try to think of this as doing everything for everybody, it won’t succeed. There’s data and there’s how people interface with that data, so we need to think about what is core data and what system can we all access in ways that are appropriate for our roles.
Billy – We’ve been thinking that there will be two to four phases or layers.
Tracy Futhey – Some things, we talk about whether they should or shouldn’t happen. Other things, it’s clear there’s a compelling reason to getting them together. This is one that needs to happen. The extent to which people have ideas for the process, that’s where we really need to go. Lots of my counterparts at other universities have been tripped up by admin systems gone bad and I don’t want to see Duke go this way. So I urge people to talk to them about your ideas related to what’s a committed activity at this point.
III. Course retention policy – Lynne O’Brien, Julian Lombardi
Julian Lombardi – We’re talking about course management materials and things that relate to the course management system that are linked to or not related to the system.
We’re really at a transition point with media and the use of media in instruction. The result of DDI has been that there has been some uptake in the use of media. It’s increasing our storage requirements pretty significantly. Mp4s start adding up and exceed what was the normal size. We’ve run into limits within Blackboard. We really view this as the beginning of a time when we’re going to see a much greater use.
So we ask, do we need a policy for course retention? The need to do so is compounded by the fact that courses very often are instances. Faculty will develop an instance, then a new instance, then a new one. It has historical purpose, possibly, and it wasn’t a problem when sizes were small, but with more system requirements, it’s bigger problem.
Lynne O’Brien – There are many course sites that haven’t been used for four or five years. So under what circumstances do we save them or delete them? One thing we want to do is have the option for people to get rid of old sites of their own. The question is, do we keep everything forever? Or keep things for smaller amount of time with an option for extending the time?
Robert Wolpert – For courses that use commercial software like the course management system, Blackboard, do we have the option of keeping things forever. Versions change over time and it’s possible that we can’t easily extract the information.
Lynne – There’s the system that aggregates the parts and displays them, then there’s the stuff. It’s easy to save the stuff, but the display changes over time. Even as we keep things forever, there’s no guarantee that the way a course looked in 1999 will be displayed the same way in version 5 in 2020. The documents and files are there. There are individual things that are interactive like chats and discussion boards – some are retained and some are not.
Julian – Our proposal is a period of three years where things are automatically continued, and after three years it’s deleted, unless there’s a request not to delete.
Question – Do you have any idea about file sizes? Are there thousands of 1K files or fewer, bigger files?
Lynne – There are lots of small ones, but the number of larger files is increasing. Right now the number of courses with a large amount of file space is small. But we see that as going to go up.
Chris Cramer – Are students allowed into a course after they’re not in the class anymore?
Lynne – It depends on whether the faculty member keeps the site live or not.
Tom Wall – The things like a course management system like Blackboard, they’re operating on their own. EReserves are cleared after six months. Some of that is law and the Patriot Act. So even if you keep your information, the eReserves won’t be available. It’s not so much the copyright, it’s the access rights.
Chris – Those rights are only for the duration of the course.
Jeff Chase – Is the volume of data growing faster than the cost of storage is dropping? Can we send faculty an email saying if you want to save this, click here?
Billy Herndon – If we don’t have a policy, at some point even if we have one, we won’t be able to access the information, depending on releases of Blackboard. It’s not just the cost of storage, but there’s overhead in the day-to-day backups of this and conversions, the length of time necessary to migrate the information. It hasn’t been an issue, but the last two semesters it has really started to grow.
Nevin Fouts – I think Jeff’s suggestion is a good one. Perhaps student access goes away unless someone asks for it to be opened up. And unless the information is accessed by the instructor for two to three years, that note comes up. Just have that as a policy. And you won’t worry about the copyright violations as much.
Tom – It would be interesting if we could take the learning objects out and keep them in a place where faculty could access them and the copyright issues won’t be there. We’d have a library with restricted access, and the course size won’t take up all that space.
Lynne – Especially for courses that have 40 sections.
Tom – You would have links, rather than 40 incidences.
Lynne – We do occasionally have students who want to go back to the course. We have students with incompletes or people who thought they did something that there’s no recrord of. We don’t routinely have people asking for that.
Tom – For a lot of people the Blackboard site is the syllabus. They want to know what the value-added items are (chats, etc). Right now, if they’re not enrolled they don’t have access to that. Both for eReserves and Blackboard items.
Julian – One criticism I’ve heard is that they take the place of textbooks. These sites are equivalent to institutions that burn the textbooks at the end of the semester because you’re eliminating information students might need as they go on.
John Board – I’m more concerned about the time it takes to recreate a course. I keep all my own websites. I’m teaching a class I haven’t taught for seven years. I lean far more on the side of preserving anything that enables faculty to develop things easier. Also I get increasingly concerned about our ability to own and control that information and not be at the whim of a company changing something.
Robert – If the material were put into secondary storage in a way that could be brought back relatively easily, that would be better. But if it just disappears after three or five years, that would be a problem. The copyrighted assets could be treated differently.
Tom – The copyright issue is mainly about the stuff you scan and change the format prior to distribution.
Nevin – A company wants to know what Duke’s doing for us. I think the trend toward things being technology advanced is good. I like the idea of extracting things to a library that can be handled. There’s a lot of awareness about IT. If you want to set things aside you have to be aware of how things are saved.
Julian – At Fuqua, don’t they give people CDs or DVDs after the course?
Nevin – In some cases. We have licenses for this time and this many students and when that ends, we have to take action immediately.
Lynne – Students have complained about not being able to access Spanish 1 before Spanish 2.
Kara Jelinkova – So we’re moving from question of delete or not to tiered storage and archival issues. The technology is available if the desire is there.
Molly Tamarkin – Can you have a time when it’s no longer part of course management system, it migrates into someone’s personal storage space?
Billy – That would be good.
Julian – The context is lost but the assets are saved.
Lynne – We need faculty to know that we use Blackboard but we see faculty relying on stuff that isn’t in the course management system, so even if we keep course forever, that stuff isn’t part of the system, so it’s not saved.
IV. Wiki and blog infrastructure – Kevin Witte and Billy Herndon
Kevin Witte – I talked at an earlier meeting about a team that was getting together to evaluate collaborative software and make recommendations about how we might deploy collaborative software in ways that would be useful. It’s been called by the misnomer wiki/blog team because those are two things we’ve been looking at. We’re really looking at a broader range – social software – looking at an environment that will allow people to collaborate. Part of the focus has been, how do we provide an enterprise-scale wiki solution.
So this team of people that was made up of folks from several divisions looked at what we needed and looked at solutions that met those needs. The team recommended Confluence and MediaWiki. MediaWiki is the engine that sits behind Wikipedia. There are tens of thousands of MediaWikis that are accessible across the Internet. Confluence OIT has been piloting for a couple of years now. It’s been popular and was saturated with users when it went live.
The Internet Framework Services team took a look at both and figured out which is a better fit for the institutional needs. We looked at them for a couple of months – looking at code and function, benchmarking what we were trying to do against other institutions. Looked at what N.C. State is doing, where they had implemented an instance of MediaWiki for everyone on campus. People on campus came here and talked about it. We also looked at Confluence out in the world in addition to OIT. What we found when we did an inventory was we’ve got instances at Stanford, MIT, Yale, Cornell, Brown, USC, Berkeley, also a number of consortia using Confluence as their primary wiki – Internet2, Sakai, JSIG.
We had a pretty good set of functional technical requirements. A couple included scalability, also making sure that we had the ability to grant access at a very granular level. We looked at the tools from a number of perspectives. Some of us were leaning toward MediaWiki because it’s open-source, but it turns out that MediaWiki has a very rudimentary access control capability.
So we decided Confluence provided, out of the box, a technical platform we would get today and get something up and running by the spring. So we’re on course for that, along with all of the policies and procedures and all the communication.
Question – When is spring?
Kevin – The thing is that we have been piloting it for a couple of years. We know it’ll integrate with webauth.
John Board – If I put stuff in the wiki and Confluence goes belly-up, how is my stuff being saved?
Kevin – Everything we’re storing is stored in an oracle database. Also, we can purchase a license that gives us the source code as well for future reference.
Lynne O’Brien – Will the authentication strategy allow non-Duke people to have access?
Kevin – We have to have anonymous access in places where that’s necessary. We can also lock down any given space and require folks with a NetID at minimum. Again we have a refined ability to lock things down at the space level and the page level.
Robert Wolpert – We don’t necessary need one system. Every year we have half a dozen times when people all over have authenticated access to information for a fairly short time. It has to be easy, quick, something a local system can bring up without a lot of training. So do we also need some cheap, quick grassroots system for that?
Kevin – Good point. We also made it clear that a couple of products out there that were good for individual or departmental wiki space. We have a couple that we evaluated and one thing we talked about doing was having a way to integrate that with our local infrastructure when that makes sense. So we need to know if we need a separate system that can do that.
Molly Tamarkin – I don’t know why Confluence can’t do that.
Chris Cramer – We don’t want just anonymous or NetID, but cross-institutional authentication.
Molly – Also, being able to export content into another format so we can take documentation and get a PDF version.
Mike Pickett – One thing that’s come up is whether we can create an infrastructure with ID management that would allow guests.
Kevin – Some consortia we’re looking at are doing that with Shibboleth.
Lynne O’Brien – We also want access for people from the community who want to comment.
V. RENCI, the Renaissance Computing Institute, update – Jeff Chase and Julian Lombardi
Jeff Chase – The mission of RENCI can be found at www.renci.org. It got started a couple of years ago in Chapel Hill. It’s still growing into its mission and trying to figure out how to serve that mission. RENCI now has a recurring funding source from the state. What’s going on now is they’re in the process of setting up collaborations, defining what kinds of partnerships should take place and how that partnering should work.
As part of that Julian and I have been tasked with representing Duke in discussing RENCI. We’ve been negotiating with them to have them make some investments in Duke on their side.
Julian Lombardi – The specific amount isn’t clear, but it’s bigger than a breadbox. They want Duke to be a RENCI engagement site. That means 2,500 to 3,000 square feet outfitted by RENCI , furniture, equipment, etc., based on the nature of potential interaction between RENCI and other institutions. Other centers are in Charlotte, Asheville, UNC, N.C. State. They’re components of a metafacility for the state, allowing connections between researchers and each other at other institutions.
John Board – What about connections within the university itself?
Julian – Yes.
So RENCI has been exploring the idea of a Duke engagement center and Jeff and I have been working with them. We have received approval from RENCI to implement an engagement center and they’re footing the bill for it.
We are presently looking at the space on campus – the first floor of the Telcom building. We have a bunch of empty offices there and the university operators. We’re looking at an upfitting of that space, which is more than 2,500 square feet.
RENCI is now willing to pay for demolition and renovation to create a state-of-the-art meeting and research facility that would contain high-end equipment of an unspecified nature. We are in the process of establishing a mechanism by which to specify that.
So we submit to you this floor plan from the N.C. State facility that has just been finished and is being outfitted. It has a display wall to interact with other centers and people across the world. It also has space for RENCI staff.
Jeff – There are several positions and we’re interviewing. These people are employed by RENCI but housed at Duke. Their mission is to fulfill the RENCI mission in a way that fulfills missions of both entities. These people will have specific technical skills. They will be tasked with operating the equipment and being liaisons with RENCI. They’re a resource for collaborations with RENCI .
John – What is the relationship between RENCI and CSEM.
Jeff – CSEM looks inward, within the university. RENCI serves the state. What they provide is a path for researchers within Duke to engage with a broader audience. To some degree both are in the business of providing tools to the Duke community. The relationship is yet unspecified. Duke really needs to control the resources that Duke requires. This is icing on the cake. We do not want to outsource the resources of CSEM. There’s clearly a link between these two things.
Robert Wolpert – Is teleconferencing viewed as an important aspect of this?
Julian – Yes.
Mike Pickett – The visualization display wall sounds like what Nevin has set up with his global conferencing. Can we get that out of this?
Julian – Yes. Telcom is in a good location and all of this stuff will be available for faculty, researchers, students. They want this to be an attractor.
Nevin Fouts – In terms of visualization application experts, would some people paid by Duke be in there of that ilk?
Jeff – Part of this relates to CSEM. One reason why it’s at Duke is because Duke has visualization capabilities. We haven’t yet figured out what the relationship between our visualization people and their visualization people will be.
David Jamieson-Drake – What do they want from Duke?
Jeff – They want it to be an attractor for researchers, they want collaboration between Duke researchers and others throughout the state. They’d also like to be able to link up to a K-12 aspect of their mission. They want to attract more federal funding. All would be success measures for the center. Certainly one of the goals will be for the center to be self-sustaining at the end of three years.
Julian – After three years if the funding doesn’t exist, the center would still be there but we’d have to find a way to fund it.
Mike – Figuring out the process of integrating CSEM with this will be tricky.
Molly Tamarkin – Are you ready for faculty projects? What’s the timeline?
Julian – We aren’t ready for that. We haven’t figured out the advisory structure, etc. Right now you could talk to either Jeff or myself about it.
David – There’s a little bit of “if we build it they will come” about this. Have you done any assessment of whether people will make use of this?
Jeff – No. That’s an issue with CSEM as well.
Mike – Having watched supercomputer centers come and go and state funding whims come and go, perhaps we could think about, if they build it, what will happen if RENCI goes.
Jeff – There’s some issues with that. Models are still evolving. I think the important thing is we should not build dependence until the resources are in place.
Nevin – We should assume the resources will not be there. We should visualize at least three things that Duke researchers would do in this space. If we can’t cast that out and ask if you could do this, this and this, then it will be RENCI-driven instead of Duke-driven.
Jeff – I would say that there are some researchers that will engage with RENCI in some ways. It’s not trouble-free because their mission is evolving.