Monday, December 5, 2011

Input sought on long-range IT plan

My office is developing a long-range plan for information technology for the university. A draft set of recommendations has been developed and I would like to share it with the community. 

Three town hall meetings have been arranged to give you the chance to hear some of the details and offer your input:
  • Tuesday, Dec. 13, 2 - 4pm, Augustana, room C167
  • Wednesday, Dec. 14, 10am - noon, Council Chambers, University Hall
  • Thursday, Dec. 15, 1 - 3pm, Council Chambers, University Hall

We also welcome your input by email at vpit(at)ualberta.ca or on our web page

Thank you,
Jonathan Schaeffer
Vice-Provost and Associate Vice-President (Information Technology)

8 comments:

  1. I'd love to see the creation and maintenence of Laptop work labs; Essentially labs stripped of machines, with just raw DVI/VGA connectors and great-big monitors (24" + ). As many students now take their own laptops from place to place, and do the majority of their work on them, the extra screen real-estate would be beneficial, while also reducing the costs of maintenence in those labs.

    ReplyDelete
  2. Excellent suggestion! We have been thinking along those lines, so expect to see some changes down the road.

    One issue is that not all students have laptops, so we have been considering a setup where, say, 20% of the spaces in a lab have computers and 80% have screens and plug-ins.

    ReplyDelete
  3. I don't have time to attend your town hall meetings (such face-to-face meetings are apparently "Parent generation" according to one of your documents): please provide a web link to the draft recommendations. All I could find on your website was a ppt presentation with no details, and a 7-page general document ("The IT Campus of the Future: The Mobile, Connected Community"), which also has no details.

    In the ppt slides, you say:

    "Doing what we’ve done in the past is not good enough....
    Need to do less with less"

    That may be the reality, but it is hardly an aspirational starting point.
    Jeremy Richards

    ReplyDelete
  4. I've heard that the University of Alberta uses Oracle as a middle tier and data tier. Have you looked towards more open-sourced options? I know support is a very important thing in an enterprise environment, but it would also be nice to further develop open source software to the University's standards as well as teaching students about the open source world. In my opinion, I don't think Oracle support is the best thing out there anyway.

    On the hardware side, have there been more natural and energy efficient ways of cooling in data centers?

    I like your idea with virtualization since it works out in the IT world so well. You can even use the decommissioned servers as a way to teach students about server hardware. I haven't done much research in it, but as a student, I feel that there isn't enough IT related courses. Especially since the DevOps concept is starting to become a reality, it would be hard to come out with just a development background.

    Regarding Matt G.'s comment about the less computers, I would like to see a higher percentage of computer labs vs laptop labs (around 60/40). Many students do bring their laptops, but there are still those who prefer taking paper notes in class and those who can't afford a laptop. Right now we don't have enough laptop spaces so it would be nice to expand that to a lab or 2, but I wouldn't go crazy about reducing the amount of computers to 20%.

    ReplyDelete
  5. The university needs to centrally support a digital archive for raw data that is collected by researchers. The Tri-Council funding agencies are moving toward transparent and open access to data developed from research that they have funded. A large proportion, perhaps most, of the data analyses that are performed during research are computational, and thus a stable archive of data is essential for reproduction and extension of the ideas that are being developed by researchers. It is also clear that the integration of teaching with research should be a central initiative of the university, and to do this effectively data need to be accessible in a common platform that allows students ready access so they can learn how to perform real research-driven analyses of real data.

    This would need to be a very large data repository, obviously, and the job of creating a database or a small number of databases, and interfaces that will allow automated and structured access to the wide variety of data types (including images) is difficult, but this kind of resource development is in fact a research program in itself, and definitely worth doing.

    If we do not this then students and researchers at the U of A will be at a long term disadvantage. If we do it soon and well we can be world leaders in both research and the integration of research and teaching.

    ReplyDelete
  6. Hi, there seems to be a mix of opinions as to whether to leave computers on or off. I come from an office in a very large company where we went from leaving computers on over weekends to turning them off, to leaving them on again, the IT department didn't get involved (they said it was up to us) so the choice depended on the manager at the time. I was surprised how IT didn't encourage turning them off or putting them to sleep, as if simply trying to avoid conflict. Anyways I think that IT professionals must communicate to their clients how they can save electricity (and help reduce environmental impact) by powering off or putting computers to sleep. It would be nice to receive some guidelines about this in our 3rd or 4th year of computing science. For example, the head of IT at a hospital could be saving 25,000 a year by encouraging people to turn off or put their computers to sleep over the weekend! (consumption of 500 CPUs running at 100W from friday afternoon to monday morning at 15 cents per KWH)

    ReplyDelete
  7. I have a recent experience that is worth passing on. Our lab has many highly sensitive instruments that are not tolerant of electrical noise. Plugging in noisy equipment into hospital grade outlets should not be allowed. My solution to this technical problem was to lend the lab a new type of UPS. This particular Cyberpower unit generates its own AC from DC and filters out almost all noise. I bought the largest 1500W 10 outlet unit for a new server build ($250) and a smaller unit for home use ($110). These are excellent products and should be phased in replacing older less efficient units with older filtering electronics. Rack mounted units are avaliable for computer labs.

    ReplyDelete
  8. The idea of a database of research data may seem like a good idea at first glance, but it is completely impractical. Coming from a phys chem lab which has been active for 40 years, I can tell you that we have several terabytes of data, which we willingly pass on to those academics requesting access. If there were 1000 equally productive labs on campus, imagine the storage space required at the core facility. In addition, opening large files simultaneously on a server and connecting to a local PC would require a lot of memory in the server centre. Downloading such files would be more realistic, but the bandwidth required for a large campus would surely exceed the current allocation. In my experience, it is almost impossible to get through to the library to download publications, when large computer labs are online at the same time. Thus, a new faster university network would be required campus wide. This is expensive, but if it generates jobs ...... What about the security issues. Databases are frequently hacked from the outside by people with suspiceous motives. How many people would be required to monitor this secure resource and maintain and update it. Researchers have enough to do as it is. Our lab always used to submit data as part of an electronic manuscript submission. This allowed the publisher to forward the files to a database formerly at the Ohio State U. In time, the job of submitting the data bounced back to the labs supplying the data. That was the point at which the system broke down. Simply put, you would need to have someone in each lab who was willing to do this and that costs money. So, while there may be money for big toys, there is no money at the basic science level to hire people to do what is being suggested. The only area in which data bases have proven successful are those assembled by X-ray crystallographers. Databases exist to store X-ray crystal structures of both organic and inorganic molecules. Many of these databases can be accessed free of charge over the internet with specialized software (also free). Centralizing everything produced by the university would ultimately lead to the need to generate funds to maintain it. That is nice for a business, but does not lead to scientific progress and so is self-defeating.

    ReplyDelete

A moderator will review your comment before it is posted to ensure it does not violate the community guidelines. Be aware that at this time you are posting to an externally-hosted website. You have no more privacy protection than you would posting to any other site on the Internet.