MIT Faculty Newsletter  
Vol. XVIII No. 5
May / June 2006
Meritocracy and a Diverse Faculty
A Brief History and Workings
of the MIT Corporation
Committees of the Faculty:
An End-of-Year Recap
Lippard and Sharp Awarded
National Medal of Science
Energy Research Council and Forum:
A Major New Institute Initiative
Efficient Use of Energy:
Part of MIT's New Energy Initiative
Fueling Our Transportation Future
Lighting a Fire in MIT's
Undergradute Education
Some Thoughts on the Arts
Reflections on the "Visualizing Cultures" Incident
On the "Visualizing Cultures" Controversy and its Implications
Communication Requirement
Evaluation Process Begins
A Modest Proposal:
A Dental Insurance Plan for All Students
New Resource on Faculty Website:
"Current Practices"
"Soft Skills" to Help Avoid the "Hard Knocks"
Computer Space Planning for MIT
Tops IT-SPARCC's Priority List
Seniors Report Increased Satisfaction
with Faculty Interaction
Smart Buy Purchasing Initiative
Primary Form of Support
for Doctoral Students
Printable Version

Computer Space Planning for MIT
Tops IT-SPARCC's Priority List

Jerrold M. Grochow

Use of information technology has become one of the most important factors in the success of academic, research, and administrative enterprise at MIT. A 2002 external review committee recommended that MIT appoint a strategic coordinating body to develop better Institute-wide processes for planning and resource allocation for all aspects of IT. In response to this recommendation, the Provost and Executive Vice President created the Information Technology Strategic Planning and Resources Coordinating Council (IT-SPARCC) to advise them on policy and priorities for IT infrastructure and services. IT-SPARCC members include faculty and senior administrators, and is chaired by me, as Vice President for Information Services and Technology (IS&T).

IT-SPARCC has been meeting for the past year to review reports and proposals on issues prepared by task forces of the Council, by staff, or by other organizations within MIT with focus on policy and recommendations with broad impact across the Institute. Topics that the Council considers include:

  • General review of the provision of IT and related services at MIT
  • Charging mechanisms for IT services
  • Strategies for allocation of the Institute general budget across academic, administrative, and research computing
  • Space allocation for IT
  • Review of priorities for computing infrastructure, academic computing, and administrative IT

Recommendations from IT-SPARCC feed directly into the decision-making process for capital allocation, annual budgets, and rate structures.

IT-SPARCC members include both faculty and administration, by design overlapping with members of the Academic Council, MIT Council on Education Technology, and Administrative Systems and Policies Coordinating Council. Faculty members include Dean of Science Robert Silbey, Dean of Undergraduate Education Daniel Hastings, Hal Abelson, Nancy Leveson, Bob Redwine, and Associate Chair of the Faculty Bruce Tidor; Professors Steven Eppinger and Lorna Gibson served on the Council during academic year 2004-5. The Council’s charter and full membership are available at

Top Priority: Computer Spaces

Over the past year, IT-SPARCC has discussed a variety of topics, and currently has computer space issues as its top priority. After an initial round of fact-finding and discussion, the Council endorsed the idea of a multi-tiered solution for critical IT space needs. Recognizing the complexity of this idea, the Council then commissioned the Computer Space Task Force (CSTF) to focus on facilities for server management and high-performance computing (HPC). CSTF membership includes many faculty and researchers, as well as IT staff from various departments. The CSTF’s charter and membership are available at

At the present time, large-scale computer servers are pervasive throughout the campus because of the relatively low-cost for cluster and server hardware.

It is estimated that administrative and HPC computers are housed in over 100 sites throughout the campus, most under faculty control.

The vast majority of these sites are not properly designed for data center usage and as a result, deans and laboratory directors are under increasing pressure to renovate many localized prime campus space locations for similar types of infrastructure for computer machine rooms. The IS&T Department also provides 4000 sq ft of space and facilities for computer server and HPC co-location management, but that is running to capacity. The opportunity cost for other more pressing academic and research uses of on-campus space coupled with the large costs of individual local space changes is prompting the Institute to seriously examine this computer space issue.

Back to top

In June of 2005, the Council endorsed a set of recommendations to the Provost and Executive Vice President that focused on the issues of computer space.

  1. That all Institute building projects (new and renovation) require planning and review for current and future computing needs for expected tenants as part of the Institute’s capital approval process. Funding considerations should be part of this process.
  2. That the Institute plan for and provide economic and scalable computing, communications, and "co-location" spaces (a secure site with good environmentals for housing servers and related equipment) to meet the current and future requirements for computing and communications hardware and related equipment.
  3. That the Provost and Executive Vice President use Institute funds for co-location where possible and agreeable to all parties. Such funding should apply to requests for computer infrastructure space, power, cooling, and network connectivity in co-location spaces that have been specifically outfitted for these needs. Institute funds should be considered for individual DLC sites if the DLC and IS&T jointly present to the Provost and Executive Vice President that the Institute co-location cannot meet the needed computer hardware infrastructure requirements.
  4. That all DLCs identify computing needs and requirements in conjunction with research proposals and annual plans, taking an approach similar to identification of other capital needs. IT-SPARCC will review this information in aggregate to ensure that IT considerations feed directly into the decisions for Institute capital allocation, annual budgets, and rate structures.

Researchers in many departments need space and facilities ranging from housing a single rack of computer equipment to as many as 40. Some faculty want space that allows easy access for themselves and students; others are able to manage their computers remotely. Some faculty are looking for basic facilities (power, cooling, and Internet access) while others want facilities with backup power, redundant cooling, or other aspects of high-reliability computer centers. Some faculty manage computer equipment using student labor, while others are looking for various levels of services from IT professionals.

Recognizing the diverse needs of the faculty for different levels of infrastructure support in this area, the CSTF is considering a number of campus and remote locations.

The task force has identified several key cost categories impacting the consideration of any options for computer space. These include the costs of:

      • Building construction and renovation
      • Network connectivity
      • Capital equipment and IT infrastructure
      • Operating costs (labor and consultants)
      • Lease costs (if applicable)
      • Power costs

From an analysis of its preliminary findings, the CSTF has observed that power for a data center/co-location site is the most expensive cost factor over the lifetime of the facility (it is estimated that MIT currently spends almost $2M per year for electricity to power and cool HPC sites across campus). Because the cost of power is similar among campus and Boston/commuting locations, the total cost of ownership over the lifetime for any such data center and HPC co-location facility are fairly similar. One would have to identify a remote location with substantially lower power costs to significantly impact the total data center/co-location costs.

Based on its findings to date, the CSTF has concluded that any solution to satisfy MIT faculty, researchers, and administrators may need to include a combination of on- and off-campus space, some administered locally and some centrally. The success of the centralized data center/co-location approach will depend upon adequate planning, funding models that provide incentives, a plan to provide excellent equipment service to faculty who would locate their computer hardware at such a site, and close cooperation with faculty, department heads, and administrators in the DLCs.

As part of this planning and fact finding process, IT-SPARCC and the CSTF are keenly interested in receiving feedback from faculty and researchers about their computer space IT needs and suggestions going forward. The CSTF and the Council welcome all faculty suggestions and feedback. The Council will delve into a variety of issues related to IT planning on an Institute-wide basis and will make recommendations to the Provost and Executive Vice President. MIT relies on well-run, cost-effective server management and high performance computing to do all the things it does so well.

In addition to talking with any of its members, you can also provide feedback to IT-SPARCC by:

Back to top
Send your comments