The CIT Public Computing Labs have been supplying computing .._1_

Document Sample
The CIT Public Computing Labs have been supplying computing .._1_ Powered By Docstoc
					Cornell Public Computing White Paper:
 Cover Page with an Interesting Title
            Submitted January 31, 2003
Table of Contents:

EXECUTIVE SUMMARY: ..................................................................................... 5

INTRODUCTION: ................................................................................................. 6

THE CURRENT SITUATION: ............................................................................... 7

The Cornell Public Computing Landscape: Fall 2002 .............................................................................. 7
  Statistical breakdowns: .............................................................................................................................. 8

Other Cornell Landscape considerations (2002): ...................................................................................... 9

CIT’s role in Cornell’s Public Computing Landscape 2002: ...................................................................10
  The mission of CIT Public Computing .....................................................................................................10


THE NEXT 5 YEARS: ......................................................................................... 11

Expected Changes ........................................................................................................................................11


ANALYSIS: ........................................................................................................ 12

Issues: ...........................................................................................................................................................12

Options and Scenarios: Overview ..............................................................................................................13

Options and Scenarios: In Depth ...............................................................................................................13


THE COMMITTEE’S RECOMMENDATIONS: ................................................... 14

PUBLIC COMPUTING 2008: (WILL NOT BE IN THE NEW VERSION) ............ 15

APPENDICES: ................................................................................................... 25




                                                                                                                                                                  2
Appendices:
   I. Committee Membership
   II. Methodologies
  III. Cornell Public Computers Detailed
  IV. Focus Groups Data
     a. Faculty Users Discussion Guide
     b. Faculty Non-users Discussion Guide
     c. Undergraduate and Graduate Student Users Discussion Guide
   V. WebSurveyor Survey document
  VI. E-Mail Survey Results
     a. Full Data
     b. Freshmen only
     c. Sophomores only
     d. Juniors only
     e. Seniors only
     f. All undergraduate students
     g. Graduate students only
     h. On-campus students only
     i. Off-campus students only
     j. All computer owners
     k. Desktop computer owners
     l. Laptop computer owners
     m. Computer non-owners
     n. Frequent/Often/Occasional CIT lab users
     o. Rarely/Never CIT lab users
     p. Users with PC preference
     q. Users with Macintosh preference




                                                                    3
VII. Lab User Survey Results
    a. Full Data
    b. Freshmen only
    c. Sophomores only
    d. Juniors only
    e. Seniors only
    f. All undergraduate students
    g. Graduate students only
    h. Non-student respondents
    i. On-campus students only
    j. Off-campus students only
    k. All computer owners
    l. Desktop computer owners
    m. Laptop computer owners
    n. Computer non-owners
    o. Frequent/Often/Occasional CIT lab users
    p. Rarely/Never CIT lab users
    q. Users with PC preference
    r. Users with Macintosh preference
VIII. Development of the CIT Public Computing Mission
 IX. CIT Public Labs Class sessions and hours summary data FY01 – FY03




                                                                         4
Executive Summary:
Nothing here yet, but should be a fairly sparse list of recommendations.
After carefully considering the issues, it is the opinion of the Committee that CIT should
strongly consider the following:




                                                                                             5
Introduction:
The CIT Public Computing Labs have been supplying computing resources to the Cornell
community for several years. The original labs consisted of dumb terminals connected to
the Cornell mainframe systems. As time passed and computing technology advanced, the
labs evolved to include first a mix of dumb terminals and Macintosh computers, and later
to the current mix of Macs and PC systems.

In 1993, the labs supported 7 computer labs (Upson, Carpenter, the Computer Science
Unix Lab, Sibley, Uris Library Tower room, MVR, and Uris hall), providing
approximately 205 computers, and employed 9 full time staff members.

By the fall of 2001, the number of labs had increased to 15, and 14 unique Bear Access
Kiosk locations had been added. Combined, the labs, kiosks, and training facilities
provided approximately 432 computers supported by 5 full time staff. 364 of those
systems were available for public use, with the remaining 69 systems in facilities
reserved for training.

In addition to the CIT labs more than doubling in size during this interval, several other
units, departments, and colleges on campus had added new labs or enlarged existing labs
of their own. This resulted in over 462 departmental computer systems that were freely
available for anyone to use, bringing the total to over 826 fully public computers on the
Cornell campus. At least another 519 systems were available on a limited basis
depending on the users’ affiliations.

Despite this growth in available systems, the CIT labs continue to be extremely busy
during peak times, with students waiting in line for the next available system, and
demand for 24-hour computing has continued to grow1. Instructional requests have also
continued to increase in number and in complexity. The CIT labs experienced a 33%
increase in the number of class sessions and a 3% increase in the number of class hours in
Fiscal Year 2002 over the previous year. The first quarter of FY03 saw a 94% increase in
class sessions, and a 27% increase in class hours over the expected demand2.

In light of the economic downturn and workforce planning efforts, it now seems to be an
appropriate time to perform a thoughtful review of the future requirements for public
computing as we look forward to the next decade. To that end, a committee of students,
faculty, and technologists has been formed to analyze the relevant issues and produce this
white paper describing the issues, options, and recommendations that should define the
future of public computing spaces at Cornell.




1
  See Appendices IV, VI, and VII – Focus Group transcripts, and survey responses for full text of the
comments.
2
  See Appendix IX for class data from FY01, FY02, and FY03Q1


                                                                                                        6
The Current Situation:
The Cornell Public Computing Landscape: Fall 2002

A detailed search of the Cornell campus was undertaken to identify all computing
resources available to undergraduate and graduate students. These resources were
generally categorized as either fully public, or unit associated. Fully public resources
were available to any student who wished to use them, while unit associated resources
were strictly reserved for students in specific courses or majors.

There certainly are additional resources that could not be identified in the time available
to us; however, we believe that the majority of available systems have been included in
this list, and that those resources not included are primarily unit associated, rather than
fully public systems.3

 Fully Public Systems: Freely available to the entire Cornell Community
Total number of FULLY Public systems:         822
Total number of Fully Public Macintosh        82
systems:
Total Number of Fully Public Windows          740
systems:
Total number of Fully Public Linux/Unix       87 *
systems:
Total number Fully Public Productivity        693
Total number Kiosk/Task Station               129
Total number public CIT systems               364 (264 Windows Productivity, 69 Mac
(Training Lab systems not included)           Productivity, 31 Windows kiosk)

Unit specific Systems: Only available to students in specific classes, colleges, or majors.
Total number of unit specific systems:       519
Total unit specific Macintosh systems:       0
Total unit specific Windows systems:         443
Total unit specific Linux/Unix Systems:      205 *
Total number unit specific Kiosk/Task        0
Station systems:




3
    Full details of the computer resources identified are available in appendix III.


                                                                                              7
Total Systems: All Cornell supplied computer resources
Total number of systems:                   1341
Total number of FULLY Public systems:      822
Total Macintosh systems:                   82
Total Windows systems:                     1183
Total Linux/Unix Systems:                  292 *
Total number of Fully Public Macintosh     82
systems:
Total Number of Fully Public Windows       740
systems:
Total number of Fully Public Linux/Unix    87
systems:
Total number Fully Public Productivity     693
Total number Kiosk/Task Station            129
Total number public CIT systems            364 (264 Windows Productivity, 69 Mac
(Training Lab systems not included)        Productivity, 31 Windows kiosk)

*: Unless specified, assumed that all systems in labs with both Windows and Linux/Unix
were dual boot. Thus, the total of Windows, Mac, and Unix/Linux systems (1557) is
greater than the total number of actual systems of (1341).

Statistical breakdowns: (need some explanations)

% of fully Public systems:                         61%
% of fully public systems owned by CIT:            44%
%CIT Systems of total systems:                     27%
%CIT Macintosh of CIT Systems:                     19%
%CIT Windows of CIT Systems:                       81%
%CIT Unix/Linux of CIT Systems:                    0%
%Total Macintosh of Total systems:                 6%
%Total Windows of Total systems:                   88%
%Total Unix/Linux of Total systems:                22%
%Total Unix/Linux in CSUG/ECE/Accel:               100%
% Fully public Unix/Linux of total fully public:   11%
% Fully public Windows of total fully public:      90%
% Fully public Macintosh of total fully public:    10%




                                                                                     8
Other Cornell Landscape considerations (2002):

                  Computer ownership is not currently required by the University. Still,
                   survey data shows that over 90% of incoming students do bring a
                   computer to campus with them4.
                  Red Rover wireless networking is installed in several buildings on
                   campus, but is not generally available outside of those locations.5
                  Computer Science is using a new course management system beginning
                   with the spring 2003 semester. In the opinion of CS faculty, this course
                   management system will likely result in an increased use of computer lab
                   resources for on-line grading activities.
                  Computer Science has created a new 20 seat computer lab in Rhodes hall.
                   This lab is not yet open to all CS students, and is not expected to have any
                   significant impact on the CIT labs.
                  Course Exchange has been implemented as an on-line system. Some
                   students will use lab computers for this previously manual activity
                   beginning this semester.
                  Academic courses are increasingly requiring students to create web pages
                   as class assignments, resulting in an increased student demand for
                   scanners, image creation and manipulation software, and other high-end
                   computing services and peripherals.
                  Students are increasing their use of technology in preparing and presenting
                   class projects. Computer enhanced spaces for students to create and
                   present those projects are becoming more popular and thus less available.
                  The computer systems recently added by the libraries are funded by
                   donated money. When those systems are due for replacement, funds may
                   not be available, and the systems may need to be removed, or may be
                   unable to meet their current roles. The overall number and quality of
                   computer resources will diminish over time if additional funding is not
                   obtained to maintain the current computer systems.




4
    See Appendix VI section a – Random Student Survey data.
5
    See HTTP://www.cit.cornell.edu/redrover/maps for current Red Rover coverage map.


                                                                                             9
CIT’s role in Cornell’s Public Computing Landscape 2002:

The mission of CIT Public Computing is to:
      Serve the academic needs of the University by providing:
          o teaching space for Professors to conduct computer supported academic
              courses
          o access to class software provided by instructors and/or departments
          o spaces for academic group computing projects
          o access to basic computer software, hardware and peripherals
          o access to high-end computer technologies such as: laser printers, scanners,
              and other computer peripherals


Serving Cornell University’s academic computing needs is a tremendously large and
important undertaking. CIT plays a significant role in satisfying the entire University’s
requirements for academic computing in an efficient and cost-effective manner.

By providing essential computing resources to all of Cornell, the CIT Public Labs help
provide a significant economy of scale cost savings to the University. If campus
departments were to try to provide the same resources, the cost would certainly be higher,
and the efficiency and reliability would be greatly reduced as computer support would be
added to existing positions. Student access to computing resources would be
significantly reduced as departments would give priority to their own students.

The CIT Public Labs, by providing computing resources on a first-come-first-served
basis can help smaller colleges and departments gain access to technologies that they
would otherwise not be able to afford or maintain. Larger colleges may also benefit by
relying on CIT to provide overflow resources during busy times in the college’s own
labs. Beyond providing computer labs, the CIT Public Computing staff also assists other
campus lab managers by providing consulting and assistance with lab management
software, Net-Print installations and troubleshooting, and any other solutions and
technologies about which they are knowledgeable.




                                                                                            10
The Next 5 Years:
Expected Changes
Over the next 5 years, several factors will influence public computing on the Cornell
campus. While it is impossible to predict all of the changes that will affect public
computing, some items can be anticipated.

Some of the changes expected over the next 5 years:

      Required Laptop computers are NOT expected in the next 5 years.
      Red Rover expansion plan in the next 5 years: with new lower rates in effect
       beginning July 1, 2003, Red Rover is likely to go through a rapid growth phase.
      The Original Mann building will be undergoing renovations beginning in the
       summer of 2003. A new lab (80 seat) is expected to open in the renovated space
       in the fall of fy06(?). At this time, we anticipate mix of standard lab and laptop
       loaner systems to make up the 80 units.
      West Campus Initiative impact on Noyes, etc.: undetermined at this time. The
       current Campus Life plan is to close the Noyes computer lab, and to open a new,
       smaller lab in each of the 5 residence houses.
      Increase in integration of computing and teaching/learning.
      Increase in on-line administrative resources for students adding to Colts, course
       exchange, etc.




                                                                                        11
Analysis:
Issues:
Although Cornell supplies a great deal of computer resources, focus group and survey
data show that the students and faculty are not completely satisfied with the results. In
order to more fully meet their needs and expectations, some changes to the current CIT
services must be considered.

Sufficient, appropriately configured computer classroom space
Sufficient student access to computer systems for class assignments
Improved student access to class software
Technology enhanced collaborative workspace
Access to high end computing resources – Videoconference, multimedia production, etc.
Technology training for staff, faculty, and students
Technology training spaces
Sufficient, centralized, secure, backed-up storage space for student files
Adequate wireless network coverage on-campus
Space to create and maintain additional academic instructional computing labs
Macintosh/PC ratio – what is the proper mix?
Improved support for laptop users on central campus.
Standard avenues for departments and Colleges to negotiate with CIT.
Improved leadership and coordination of public computing issues.




                                                                                            12
Options and Scenarios: Overview
Bulleted list of options here.


Options and Scenarios: In Depth
Detailed analysis of each identified option.




                                               13
The Committee’s Recommendations:
Changes CIT can consider to better meet Cornell’s needs:
More detail to come:
    CIT should provide general computing resources, allowing departmental labs to
      handle the specialized needs of their units.
    CIT must provide more central organization for public computing
           o A public computing roundtable for exchange of information
           o A designated person to organize all public computing initiatives with
               campus
           o Hardware and software solutions for the campus
           o Better consulting services and support for unit labs
    Access to more high technology for students to relieve unit labs from creating
      single specialized systems.
    CIT should create a standard method for accepting requests for new labs, or for
      taking control of existing labs




                                                                                       14
Public Computing 2008: (Will not be in the new version)
A look at how Public Computing at Cornell can be modified to meet the needs of the
Cornell campus.

Careful review and thoughtful consideration of the focus group and survey data
inevitably leads to the conclusion that the single greatest problem facing the CIT public
computing labs is excess demand on the current resources. Comments regarding
overcrowding and long waits for computers are directly due to excess demand.
Comments proposing a reduction in Macintosh computers so that the number of PC
systems can be increased, requests for increased access to loaner laptop computers and
complaints about non-academic use of lab computers, while perhaps less obviously so,
are also demand related.

This excess of demand has caused numerous problems. In order to supply enough
computing resources to meet the student demands, the labs have resorted to adding as
many computers to every available lab as possible, making the labs crowded and
uncomfortable, and poorly suited to instructional uses.

In order for the labs to efficiently and effectively function as the academic resources that
they are intended to be, steps must be taken to balance the supply and demand. Once an
appropriate balance is attained, those labs that are designated as instructional resources
can be rededicated to academics by ensuring that the room layouts are appropriate for
teaching and learning.

Balancing the demand on the currently available resources requires either an increase in
the number of available resources to meet the existing demand, a decrease in the demand
to allow the existing resources to meet the need, or a combination of the two approaches.

Options and alternatives for reducing excess demand in the CIT public labs:

      Increase supply of Public Computers
           o Use a few well placed large labs
           o Use available small public spaces
                   Hallways
                   Foyers
                   Small rooms, etc.
           o Use a combination of large and small spaces

      Decrease the demand for public computers
          o Require laptop/portable ownership and use
          o Encourage laptop/portable ownership and use
                   By recommendation of approved and supported systems
                   Through lease to own options


                                                                                          15
          o Restrict use of the labs to only those activities that are consistent with the
            mission of the labs.

Increased Supply Options:

      Increase the supply of Public Computers:
           o By creating a few well placed large labs
                    The most easily supported option, but space is at a premium on
                      campus, especially in the central areas where more computer labs
                      would do the most good.
                           Requires an increase in funding to purchase, secure, and
                              support the additional systems.
                           Requires new furniture purchases.
                           Increase in costs to provide network connections.
                           Requires additional staffing to provide support.
                           Can create more instructional space for academic needs.
                           Advertising locations and getting students to use the new
                              labs can be an issue.
           o By using available small spaces such as hallways, foyers, small rooms,
              etc.
                    Solves the space problem to some degree, so is a more likely
                      scenario to enact.
                           Requires increase in funding to purchase, secure and
                              support the new systems.
                           Requires new furniture purchases.
                           Increase in costs to provide network connections.
                           Requires a large increase in staffing to provide support due
                              to the widely distributed nature of this solution.
                           Problems will go unnoticed for a longer time due to
                              inability to monitor systems often enough.
                           Large increase in security issues, vandalism, theft, and
                              other abuses.
                           Large increase in complexity of tracking and managing
                              systems, repair status, warrantee, etc.
                           Greater initial cost for electrical and network installations
                              likely.
                           Approaching the Student Assembly resolution requesting
                              an e-mail kiosk in every campus building.
                           Most use of these distributed systems would likely be non-
                              academic.
                           Advertising locations and getting students to use the new
                              systems can be an issue.
                           Tracking usage to see if systems are located in useful areas
                              is a problem.




                                                                                         16
           o By using a combination of a few large lab spaces and several smaller
             available hallway, foyer, etc. spaces.
                 Combines all of the good and bad aspects of the previous
                    scenarios.


Analysis: Increasing the supply of available public computers can provide some relief
for the overcrowding of the existing labs.

Placing a few large labs in strategic locations is the most easily supportable option, but is
the least likely to be enacted. Obtaining space that is adequately sized and configured in
strategic locations on central campus is highly doubtful. Even if this could be
accomplished, determining the properly strategic locations, advertising them to the
student population, and getting students to use those new labs often takes some time, so
the full benefits would not be realized for several years after the installation.

Adding systems to any small spaces that can be obtained and configured to pass ADA
and fire code standards can be done, even with the central campus space contention.
Spaces that would typically be empty could be converted into mini labs by simply
installing the appropriate services (power and network) to the locations, and buying new
furniture and computers for them. If we assume the same number of overall computer
systems is added to central campus in both scenarios, this set-up would have significantly
greater staffing requirements due to the distributed nature of the computers. The simple
task of checking to ensure that a system is working properly becomes a huge problem in
this scenario. Any staff member responsible for distributed computers would be unable
to manage as many systems as staff in a current lab environment. This loss of operational
efficiency is contrary to the workforce planning initiative. Additionally, good customer
service would be extremely difficult to provide, as lab staff would be unlikely to be
available if a student had a problem with a computer system in any location.

Combining the two options to create a few large labs and several kiosk locations
combines the problems associated with both possibilities, and does not fully realize the
potential of either. The number of labs would be fewer, and so there would be less
additional instructional space, and the number of kiosk locations would be fewer and
students would be less likely to find those locations convenient. Customer service
problems would certainly affect the kiosk locations. Overall, the staffing support in this
model would be closer to the full kiosk model, so a disproportionate number of new staff
would be required.

In short, creating a sufficient number of new labs to deal with the problem is unlikely due
to the lack of central campus space, and adding enough kiosk locations to handle the
demand is an extremely inefficient model. It would require a significant increase in
public computers to effectively solve the issue, resulting in large numbers of idle
machines during non-peak times. No new functionality or improvements to the existing
model would be introduced.




                                                                                           17
Decreased Demand Options:

      Decrease the demand for public computers
          o Required laptop ownership and use
                   Removes choices from students
                   Increases cost of Cornell education
                   Likely to meet with significant resistance
                   Requires Red Rover improvements
                   Other infrastructure improvements recommended
                         Burden falls to students to find ways to use laptops
          o Encourage laptop ownership and use through recommended systems
                   Allows students to choose a recommended system or opt out
                   Reduced impact on tuition costs vs. required ownership
                   Should not generate significant resistance
                   Requires Red Rover improvements
                   Requires other infrastructure improvements
                         Burden falls to Cornell and CIT to create an environment
                            that encourages laptop use
                         System must benefit students or it will fail
                         Should create excitement and improve students ability to
                            complete class assignments

Analysis: If increasing the supply of public computers is not possible, or is not an
efficient use of resources, then the alternative of decreasing the demand for public
computers must be considered. In order to decrease the demand on the CIT labs, lab
users must be encouraged to seek out alternative sources for their non-academic and
general computing needs.

Requiring laptop ownership among incoming students is an option that has been used in
some other institutions and even in Cornell’s own Johnson Graduate School of
Management. This prerequisite is generally more common and more readily accepted in
the graduate school ranks, and is likely to create significant resistance from the
undergraduate students. The entire cost of purchasing laptops computers for all incoming
students would lead to increased tuition, creating an even higher financial barrier to some
students. With current computer ownership among Cornell students at well over 90%,
the benefits of a required laptop program are overshadowed by the negative impact it
would have on tuition and student satisfaction with Cornell and CIT in general.

Creating a program of recommended laptop systems that students can choose to purchase
would seem to be a more prudent course of action, especially when combined with a set
of initiatives aimed at improving the campus infrastructure to better support portable
computing.

Naturally, any initiative aimed at increasing ownership of laptop computers, and
increasing usage of those laptops in the central campus areas in lieu of computer lab
usage will need to first identify the barriers that currently prevent students from using


                                                                                            18
their own computers. Once those barriers are identified, a series of responses may be
crafted to address and remove the obstacles.

Current barriers to student use of laptop computers on the Cornell campus include:

      Limited Red Rover coverage
      Student issues with laptop computers
          o Some students experience difficulty setting up laptops for Red Rover
          o Many student owned laptops are not currently equipped for wireless
          o Most current laptop systems weigh too much to carry
          o Laptop systems are considered too fragile to carry on campus
          o Laptop systems are too prone to theft
          o Loss or damage to laptop can lead to loss of vital work/data
      Students still need to use the labs to access course software


In order to create an environment in which those obstacles no longer exist, or are greatly
alleviated, Cornell and CIT should launch the following set of initiatives:

      Aggressively expand Red Rover wireless networking coverage.
      Create a program to recommend a selected few portable/laptop computers for use
       on the Cornell campus.
           o Include colleges in the selection process.
           o Negotiate volume prices.
           o Consider Lease to own alternatives.
      Create an on-campus service and support center for the recommended systems.
      Create an Active Directory with all Cornell Kerberos principals defined to
       provide authentication and authorization services.
      Create a central networked storage system that will allow students to save files
       and data in a secure location.
      Create a central application server to allow wide-spread access to course specific
       software




                                                                                         19
Details:
Aggressively expand Red Rover coverage: Beginning July 1, 2003, the cost of
installing and maintaining a Red Rover access point will decrease significantly due to
changes in the Network Billing structure. Additional details will be added here to show
the difference between FY03 and FY04 Red Rover costs. It seems likely that these
changes will encourage a rapid and widespread increase in available service.

Create a program to recommend a selected few portable/laptop computers for use
on the Cornell campus: In order to encourage more students to purchase and use
laptop/portable computer systems, Cornell must be able to assist students in making
purchase decisions. By creating a small list of systems that will meet the needs of
students in their academic studies, Cornell can provide that assistance. By keeping the
number of recommended systems to a minimum, volume purchasing can provide
leverage to obtain better prices and support options than can be gotten by individual
students, and support issues are minimized by standardizing the variety of hardware. It
may be possible to negotiate 4-year warrantee periods so that the systems would be
covered for the student’s entire expected undergraduate college career.

Cornell will need to consider and account for the needs of the different colleges, and it is
possible that some of the recommended systems would not be appropriate for students in
certain disciplines. This should not cause any problems as long as those additional
requirements are fully detailed in any distributed publications. For example, the
minimum recommended system may not be advisable for students in CS/Engineering, or
Art classes, but may be perfectly acceptable for all other students. Prominently noting
those exceptions in the publications would solve those issues.

An added value to students purchasing from the list of recommended systems would be
the possibility of having the systems pre-configured to best utilize Cornell’s
infrastructure. Bear Access could be installed, Red Rover could be set up, and the system
serial number could also be registered with Cornell Police.

It is important to note that these recommendations are made only to assist students and
prospective students in making buying decisions. These recommendations are not a
required laptop program, nor would students be discouraged from bringing any computer
system that they desire to campus.

An additional option to consider is a lease-to-own program wherein students could pay
monthly payments until the end of the lease period of 3 years, with the option to purchase
the computer at the end of the lease, or begin a new 1-year lease on a new system.

Create an on-campus service and support center for the recommended systems:
One of the greatest impediments to student use of portable computers on central campus
is the fear of accidental damage. Students, with so much of their academic and personal
lives stored on a portable computer, are assuming considerable risk and responsibility
whenever they carry it on campus. If a small number of systems can be recommended
for students to purchase, it becomes possible to provide improved service and support. A



                                                                                          20
small on-campus repair center with a stock of standard spare parts and a few “hot-spare”
systems to loan out to students whose computers require more extensive servicing would
allow students a greater sense of security and freedom. Since this center would only
service the Cornell recommended computers, it would also serve as an additional
incentive for students to purchase those recommended systems. An increased volume of
purchases through that program would generate increased leverage for Cornell in
obtaining better pricing and warrantee coverage from the vendors.

This service center could be set up as a Cornell initiative, or outsourced to the Computing
Center, or another service vendor.

Create an Active Directory with all Cornell Kerberos principals defined to provide
authentication and authorization services: It is assumed that the majority of systems
that students will purchase and use in the near future will be predominantly Windows
systems. Any authentication and authorization service will need to support Windows
systems, and as many other systems as possible. Active Directory can authenticate and
support Windows clients, Macintosh OS X clients and Unix/Linux clients. Unlike the
current Kerberos/SideCar system, authentication to Active Directory is at login,
providing system security, and allowing for automated access to networked storage and
other similar services.

In order to create an Active Directory system that can be used to authenticate Cornell
users, a link between the existing Kerberos database or LDAP database and an Active
Directory domain must be established. This type of system has been successfully
implemented at the University of New Hampshire. A copy of the UNH presentation at
the SIGUCCS 2002 fall conference is available online at http://at.unh.edu/siguccs/ .

If another authentication system is available that will allow the same functions, that
alternative can and should be investigated as well.

 See the Mac OS X server Admin guide, Page 639 for Mac OS X information.
http://www.apple.com/server/pdfs/MacOSXServer_AdminGuide_121902.pdf

See http://online.securityfocus.com/infocus/1563 for Unix/Linux information.

Create a central networked storage system that will allow students to save files and
data in a secure location: Central, secure storage for computer data becomes an
increasingly vital service as more of the Cornell community become users of portable
computers. Backing up a laptop or other portable system, especially one with large data
files, can be an extremely difficult task since many backup systems require access to the
data during times when the computer may not even be connected to the network. Rather
than rely on users to create and maintain backups, or rely on them to ensure that their
computer is connected to the network and turned on at the time that a backup for the
system is scheduled, central storage space can be used to securely store vital files.
Central storage space also allows users to access their data from any other computer
whenever it is needed.



                                                                                         21
Breaking the connection between the hardware and the data will ensure that a dropped or
otherwise damaged laptop will not cause any significant loss of work or data. Like the
proposed on-campus service and support center, this initiative will allow use of portable
computers without fears of catastrophic loss in the case of a simple hardware failure or
accident. Computers purchased through the recommended systems program could be
pre-configured to map the default save location (My Documents folder on a Windows
computer) to the user’s server share. Documents created on those machines would
automatically be saved to the server by default, without requiring any additional user
intervention.

Create a central application server to allow wide-spread access to course specific
software: Even with each Cornell student bringing their own computer system with
them to classes, access to course software remains as an issue. The software that they
would need to use for coursework is too expensive for the university to purchase and
supply to each student, and also too expensive to force the students to individually
purchase.

Currently, the CIT and departmental labs fill this role by providing software either on a
number of systems equal to the number of licenses, or by using KeyAccess or some other
license enforcing software to restrict the number of copies of a program that can be used
simultaneously. Clearly, this model will cause students who need to access any course
software to come into the existing labs for that purpose, and once a student is forced into
the labs for any reason, that student will not be getting any value from carrying a portable
system.

In order to make course software more widely available, a secure central application
server should be installed. Since nearly all of the course software that the CIT public labs
are currently supporting is Windows based, a Citrix server or other competing solution
should be considered. In addition to making course software more available to students
on central campus, software would be available to students using other operating systems,
or in dormitory areas, or any off-campus location with a network connection. Students
with older, less-capable computers would still be able to run any server based courseware
applications as well as students with newer high-end computers since the processing is
done on the server rather than the client computer. If it became necessary in the future,
similar application servers could be installed to support Unix/Linux applications.




                                                                                         22
Summary:

In order to allow the CIT labs to focus on their primary mission of supporting the
academic needs of the university, steps must be taken to alleviate the overuse and
crowding of the existing labs. Several different options have been evaluated, but only
one of those options appears likely to have a long term positive benefit to Cornell.

Adding more resources in order to satisfy the demand will result in over supply due to the
natural flow of students across the campus. Peak demand times vary by the campus
location, with central campus peaking in the late morning to late evening hours, and
residential areas peaking in the early mornings and again in the late evening and through
the night. Enough computer systems in any given area to cope with the peak demand
hours would result in idle resources during the several non-peak hours. While this is a
necessary condition in order to provide for academic and safety net needs, it becomes an
increasingly inefficient model.

Required laptop computers are likely to meet with strong resistance from many groups,
and are considered by many to be contrary to the university’s equality of opportunity
stance. Any program that requires additional expenses such as laptop computers will
either increase tuition costs, or remove funding from another important resource.

Cornell is also a place where individual freedom and choice is highly regarded, and
removing the choice from students by mandating a computer ownership program is
contrary to that fundamental nature.

So, the remaining option of creating an environment that encourages the use of portable
computers while still allowing individual freedom of choice would appear to be the best
available option. Creating that environment will require a significant commitment to
building a new infrastructure that will make the use of portable computers so alluring that
students will routinely carry their own computers with them. The immediate needs for
creating that atmosphere are:

      Expand Red Rover wireless network coverage on campus
          o Red Rover service should be expanding in FY04 due to reduced cost
      Recommend a small number of portable computers for students to purchase
          o Negotiate volume pricing and extended warrantee coverage for those
              systems
          o Register all systems purchased through this program with Cornell police
      Provide an on-campus service and support center for those recommended systems
      Provide each student with secure server-based storage for saving work
      Provide a secure central applications server for course software
      Create a Cornell Active Directory with Kerberos principals to provide security
       and appropriate access to Cornell systems

With this improved infrastructure, the CIT labs would be less stressed by student use, and
could then reduce the number of computers in instructional spaces and improve the


                                                                                         23
layout of those spaces to better support instruction. Students who do not own computers
would also be more easily able to access the resources in the labs due to the reduced
demand. With a reduced number of computers to purchase and maintain, and with less
crowded lab spaces, the CIT labs would also be better able to provide and support
additional high-end computing services to the campus community.

All of these initiatives will have an impact not only on the CIT labs, but on all academic
computing at Cornell. This set of proposals would create a Cornell that would allow any
member of the Cornell community to access their data and course software from
anywhere, at any time, providing immense benefits to the ability of the University to
achieve its academic mission, and benefits to the business of managing the University as
well.

The CIT public computer labs will continue to provide essential services, but those
services would be more focused on meeting academic needs and providing access to
leading edge technology. Eventually, the labs will become central service centers where
students meet for classes, to collaborate on projects, to print, and to use computer systems
that provide video conferencing, multimedia production and editing tools, and other
advanced services to enhance their education.




                                                                                         24
Appendices:

   I. Committee Membership
   II. Methodologies
   III. Cornell Public Computers Detailed
   IV. Focus Groups Data
        a. Faculty Users Discussion Guide
        b. Faculty Non-users Discussion Guide
        c. Undergraduate and Graduate Student Users Discussion Guide
   I. WebSurveyor Survey document
   II. E-Mail Survey Results
        a. Full Data
        b. Freshmen only
        c. Sophomores only
        d. Juniors only
        e. Seniors only
        f. All undergraduate students
        g. Graduate students only
        h. On-campus students only
        i. Off-campus students only
        j. All computer owners
        k. Desktop computer owners
        l. Laptop computer owners
        m. Computer non-owners
        n. Frequent/Often/Occasional CIT lab users
        o. Rarely/Never CIT lab users
        p. Users with PC preference
        q. Users with Macintosh preference




                                                                       25
     I. Lab User Survey Results
           a. Full Data
           b. Freshmen only
           c. Sophomores only
           d. Juniors only
           e. Seniors only
           f. All undergraduate students
           g. Graduate students only
           h. Non-student respondents
           i. On-campus students only
           j. Off-campus students only
           k. All computer owners
           l. Desktop computer owners
           m. Laptop computer owners
           n. Computer non-owners
           o. Frequent/Often/Occasional CIT lab users
           p. Rarely/Never CIT lab users
           q. Users with PC preference
           r. Users with Macintosh preference



Appendices Cut for this draft. Will be added later.




                                                        26