Council on Computing Meeting
Thursday, December 17, 2009
Lisa Celone, James Feyrer, Mark Franklin, William Garrity, Robert Gross, Anthony
Helm, David Kotz, Steve McAllister, Barbara Mellert, Casey Murray, James Moor, Soo
Young Park, Stan Pyc, , Xun Shi, Ellen Waite-Franzen, Gail Wallin, Martin Wybourne,
James Moor called the meeting to order at 12:30 pm.
1. Review of Minutes of the October Meeting
James Moor turned the meeting over to Ellen Waite-Franzen. She drew the Council’s
attention to the minutes from the October meeting and asked if there were any additions
or corrections. James Moor pointed out the misspelling of his name. The minutes were
approved with spelling changes.
2. Update on Email and Collaboration System Selection
Ellen then turned to the update on email and collaboration system selection. Ellen
said that she did not have much additional information since the October meeting.
Microsoft will be on campus to do a presentation of the Business Productivity Online
Suite for major Mac users and for the COC. She asked if there were any questions.
Anthony Helm inquired about a timeline to upgrade email, upgrade the router/edge
and traffic shaping, with fuzzy name matching requiring trips back and forth across
the edge. Ellen said that fuzzy name matching would be handled differently with a
new system. Robert Gross asked if the switch to Microsoft is a done deal. Ellen said
that it was not. He asked if Google would be also giving another presentation. Ellen
said that they would not. The question was raised whether institutional hosting
was still being considered. Ellen said that prior study of this had shown that it would
not be cost effective. Asked about whether the experience with Exchange has gone
well, Ellen indicated that it had. Asked if there were any surprises, she noted that
people operating two calendars have had some difficulty, but that federation should
help with this. David Kotz asked if transfer would happen in stages the first year.
Ellen noted that when they put the project management team together, one of first
decisions made was to roll out the new system in stages. Computing Services would
be the first group and Jeff Horrell had suggested that the Library would be a good
early adopter. Anthony Helm asked about core functionality and whether Microsoft
has a timeline for full functionality with Apple. Ellen Waite-Franzen said that she
could not disclose a timeline, but affirmed that it would happen within the year.
Technical Services is currently testing with Snow Leopard and is not having any
issues. James Moor asked if BlitzMail folders moved over so that nothing is lost.
Ellen pointed out that when we went to the Exchange environment, email was
moved and people did not lose anything. Soo Young Park asked if BlitzMail is a
contract that ends such that we have to make a decision. Ellen noted that BlitzMail
was created and developed by Dartmouth, so there is no contract that ends. It is not
as full featured as it should be and users want more features than the BlitzMail
client can provide. Barbara Mellert pointed out that Blitz does not always talk well
with other email. When you get message from Gmail or Yahoo they come as
enclosures. If you’ve moved to another client to access BlitzMail, it is not a problem.
Asked about cost, Ellen said that transition would have a cost in terms of support,
but we would save on storage, back-up, and maintenance.
3. Strategic Budget Reductions and Investments Process
James Moor then moved the meeting to the topic of the budget reductions and
investments process. Ellen stated that Dartmouth is planning to cut $100 million
dollars in expense reductions or new revenues for fiscal years 2011 and 2012.
There are a number of areas that cross the college, and the leadership team is
looking at changes institution-wide. Last year the budget reductions were done on a
division/department basis and we did not look at making any institutional changes
that might result in savings. This year, there are some cross institutional areas that
are being reviewed to see if we can accomplish savings by making institutional
changes in how we work. These areas include compensation, facilities, research,
communications and IT.
We are looking at ways to reduce IT cost. The budget management team headed by
Steve Kadish is looking at types of projects that can result in large savings
institution-wide. We are considering consolidation of servers and other ways to
reduce expenses and use resources in more efficient and effective ways. David Kotz
asked if they are interested in suggestions. Ellen emphasized that if you do have
thoughts about reducing cost for IT or generating revenue, please offer them. James
Moor asked how Ellen sees the process working. Ellen said that suggestions and big
ideas will be run by leadership first to see if there is interest in pursuing the ideas.
Project working groups will be formed and the teams will tackle the issues and
make suggestions for change. She asked if there were other questions, comments,
or suggestions. Robert Gross said long ago, Dartmouth provided computing services,
resources, rentals, licenses for local merchants. Ellen pointed out that you start
getting into other kinds of legal rules when you are working with business. A
question was raised about The Computer Store. Ellen said that the Computer Store
is based more on service than sales, and the question could be raised whether the
sales aspect is needed. We will be looking to consolidate resources, not just
hardware, but human resources as well. We currently have a landscape where there
are groups of IT operations and support; humanities, DMS, Tuck, Thayer, library,
departments. Can some of the services be consolidated in some way?
Anthony Helm asked if outsourcing is on the table. Ellen noted an Educause article
that addressed use of crowdsourcing to increase help desk support, people helping
people. We have to rethink how support is given. Robert Gross asked if there are
different needs so that it makes sense to have different groups, for example the
library. Ellen confirmed that we have different types of support and different ties
among faculty, staff and students, and need to provide services to them. Some
people need and get more help than others. Staff often do the same type of work in
all of the areas where we have IT support. Mark Franklin pointed out that some
support in Thayer, like support for the 3D card, is very specialized. Other changes
are being considered, like using thin clients rather than PC’s and Macs that have a
five to eight year life and low maintenance in offices. The Library is interested in
4. Research Computing Update
James Moor then turned the meeting turned over to Jaime Combariza for an update
on Research Computing. Jaime reported that the public cluster for research, the
Discovery Cluster, now has 872 processors distributed among 15 stakeholders. The
main users are Jason Moore's group [Genetics], Andy Kern’s group [Biology], the
Physics group which shares resources, Chemistry users, and Thayer users.
Stakeholders are starting to replace their original compute nodes [Barrett Rogers
and Keith Paulson]. Utilization shows high loads during regular business hours.
Sixty-five scientific applications are installed and there are 115 users. The
Discovery Cluster is also being used for classes [Chem 62 and Chem 72/73] affecting
about 60 students. The cluster is working well and is highly utilized. There are some
problems with cue wait times and resulting questions of fair utilization. There is a
fee for users and grant accounts can request CPU’s for research. Martin Wybourne
asked if there are any visualization display walls being considered. Jaime said that
they have considered it but do not have funding.
Another advance in Research Computing is the increase in local research storage
[Rstor]. Currently RStor has three file servers and four disk vaults. Two file servers
are located in the Berry Machine room and two are located at the Etna Data Center
and act as backup with 30 TB capacity.
Biology is the first department to use the system. Five research groups are already
using RStor, three groups are in a test phase, and many more have been invited to
use this resource. The cost per year is $493/2 terabytes, with a minimum unit of
100 GB for $49.30/year. The resource can be used to store research data as
required by OSP and some funding agencies. Jaime was asked where funding comes
from. He said that grants fund most of this and researchers are strongly encouraged
to include a budget line for storage that includes any retention time [required or
desired]. Jaime was asked how this cost compares to Amazon. Jaime noted that
Amazon is 10 times this price; researchers have to pay for access in and out of
storage. You pay ten to fifteen cents per gigabyte per month for storage and also pay
for transfer. The question was raised about what you do if you run out of grant
money. Jaime said that if there is no money for storage, the data is taken out and put
on tape or hard drive. Offline storage is much cheaper, but it takes 2 days to retrieve.
Research Computing submitted fifteen projects to RCOS for approval in FY10. These
include replacing and maintaining servers used heavily; building spine treatment
outcome calculator; creating an electronic notebook collaboration tool to collect and
manage data from experiments and make it easily accessible to members of the
research group; work on code for Auroral Rocket Data acquisition and display
software; physics and research visualization; the Bamboo Project which is a multi
institutional effort to develop and sustain cyberinfrastructure for the Arts,
Humanities, and Interpretive Social Sciences. Project directors are working on a
final revision to a proposal to the Mellon Foundation for this. We are working with
Mary Flanagan, chair for Humanities, faculty members in the humanities, libraries,
and IT, and are planning to coordinate 2 events per year [early-mid May and in the
fall], to start a dialogue and promotion of using technology for teaching, research,
Jaime concluded by pointing out the work they are doing to increase efficiency and
provide more revenue for Research Computing. Currently, Research Computing is
paid for some services and provides many services for free. He is working to create
a Research Computing charge model in which people will be assigned a group
allocation subsidized by the college and any consulting services beyond this
allocation will be charged. In this model, there are two levels of support. Basic
support is free of charge and includes assistance for support services like installing
and testing software, compilation of research codes, limited debugging, training.
Advanced support is directed towards scientific applications used primarily by one
group and aims at improving efficiency and usability of codes and resources. This
would include programming [code optimization, code parallelization, code
development, intensive debugging]; visualization and graphics; development of user
interfaces; statistical consulting; scientific application support. Advanced support
would require budget lines in proposals. Robert Gross pointed out that if everyone
who writes a grant needs to have a line item, then services shouldn’t be free. The
question also rose regarding what would happen to support programs with
graduate students. Martin Wybourne cautioned that the model has to be careful not
to double dip. David Kotz pointed out that a cost center would have to charge
people even if they don’t have grants. Jaime acknowledged that it is a complex
model and the institution would have to subsidize. Jim Feyrer said if it brings more
money to the institution as a whole it would provide more resources, but it also
might be a way to redistribute resources. The problem is to get outside money to
grow the research group and to meet needs, not take money out of one department
and put it into another. We need to be clear about what is being accomplished:
more money or redistributing money. Anthony Helm pointed out that this might
adversely affect faculty who are not grant writers. David Kotz noted that he’s not
sure you can add an IT tax. Direct cost can be written into grants, but you can’t
write flat fees into every grant. The cost might be difficult for areas such as video
research; studio art is a good example.
Ellen asked if there were further comments or questions.
Ellen was asked about open positions in Computing Services. Ellen said that there
are still questions about the open positions. She is holding some positions open
until the budget cuts are clearer, preferring not to hire positions before layoffs.
Ellen asked that people bear with her as we figure out the bigger picture going
James Moor thanked everyone for coming.
The meeting adjourned at 2:00 pm
Recorded by Kathleen Moore