http://virtualschool.edu/nsf Updated May 7, 1998 by Brad Cox

Effects of Technological Innovation on Organizations

A GMU Proposal in Response to the 1998
NSF Knowledge and Distributed Intelligence Solicitation

Budget Explanation

We propose a three-year project to start in Spring of 1999. The budget assumes each student takes a course load of two courses/semester and that they proceed through the degree as a cohort which is usually but not always the case.

Personnel Section

Management Team: Arthur Melmed (lead), Brad Cox, Thomasina Borkman. Advisors: Haynes, Finkelstein, Fukuyama, Stowe. Consultants: Perelman
The management team provides ongoing project support and management services to the other teams.
The advisory members, which are not funded under this proposal, are the director of the MA in International Transactions degree program (Fukuyama) and the director of the Institute of Public Policy (Haynes) which houses this degree. Finkelstein and Stowe are included as a source of general guidance and advice.
Twenty days of consulting time is included for as-needed advice in the evolving relationship between education and its social environment.

Deployment Team: Brad Cox (lead), 4 Masters Students
This team provides technology deployment and support services to the faculty responsible for MAIT courses. The deployment team should be sufficient to support two new courses (O1) per semester and to provide ongoing extension and support for O2 and O3.

Development Team: Brad Cox (lead), 2 Doctoral Students
This team refines the current coordination infrastructure and supports the deployment team in applying T1 and T2 to O1 (courses) and O2 (degree). The development team will also be responsible for O3 (project) which serves as this team's testbed for new software releases.

Interdisciplinary Studies (Impacts) Team, Thomasina Borkman (lead), Thomas Ratchford, Chris Rodrigo, Jack High, Richard Klimoski, Laurie Schintler, Rajendra Kulkarni
This team will study the impacts of technological deployment by initially collecting baseline information prior to deployment (T0) and subsequently by following the T1 and T1+T2 impacts over the life of the project.

Fringe Benefits

These figures were computed automatically by sponsored programs' spreadsheet.

Travel

This assumes one trip per PI during the first year, and one trip per senior researcher and PI in the second year to attend relevant conferences. The final year includes half the second year travel to relevant conferences to present papers arising from this work.

Supplies

The bulk of this section is for computer software. The figures I've provided are estimates. Exact brands and prices won't be known until we've completed investigations already planned, independently of this proposal, for Sum 1998.

Server-side software expenses include an object-oriented DBMS such as ODI to replace the current flat-file perl database, an up to date web server such as Apache, Netscape Enterprise, or Microsoft, CORBA development environment such as Visigenic, a WYSIWG web development environment for team development such as ColdFusion and/or TeamFusion, and a web-based discussion environment such as Caucus.

Client-side expenses would include file transfer clients for each developer such as CuteFTP or PCAnywhere depending on whether Unix or WindowsNT is chosen for the server, and a WYSIWG html editor for each contributor such as Frontpage or ColdFusion.

Exact prices won't be known until Summer 98. We budgeted $8000 total, half to be spent in the first year.

Equipment

The dominant expense is for web server capable of handling students, faculty and developers, building from 100 to 400 students by the third year. The existing Unix/BSDI server will be retained and used for software testing. We've specified a capable machine (>$5000K) to provide no-excuses service quality to students and because T2 applications are cpu- and io-intensive. We've also budgeted an additional $8000 for client computers for any faculty and support staff who may not have suitable computers already.

Other Direct Costs

Most of the items are estimates based on the research team's experience with similar projects.

The major item, $44,200 ($1000 for installation and $1200 per month) for "Server Housing and T3 Backbone Access" is the rate quoted by Erols Internet Services (www.erols.com) for "colocation", which covers housing the server on the ISP's 100mbit/sec internal network, multiple-T3 connection to the internet backbone, and support desk access. The new server will run Unix or WindowsNT depending on software compatibility issues TBD. The goal is rock-solid no-excuses reliability and enough computer and network horsepower to handle heavy loads from as many as 400 students and several support staff running cpu-, io-, and network-intensive programs.

We budgeted a commercial internet provider instead of housing the server on-campus because of reliability and support differentials between commercial and academic institutions. Students have no tolerance for unplanned outages so the slightest outage triggers a deluge of calls to faculty who will often lack the technical experience to be of much help.

Experience with the Taming the Electronic Frontier course shows that most (80%) students (in the case of MA in Telecom students) already use external ISP services in preference to GMU's free but overloaded service. If the server is on-campus, the bulk of the traffic funnels through GMU's backbone connection, LAN and routers. Although service levels are improving, these are non-redundant systems with numerous points of failure. Although the most common cause of computer failures at GMU are campus-wide power outages, there is no backup generator and only enough UPS battery capacity to support orderly shutdowns.

We chose this vendor because of our experience with the virtualschool.edu server, which supports the Taming the Electronic Frontier course. This server was donated to the PI by Erols' president three years ago and is housed and connected as we've proposed for the new server. Availability and performance has been outstanding with only a single outage for the last 12 months. Although GMU's service is improving, it has averaged an outage per week during the same period and seems to be chronically overloaded.

Indirect Costs

These figures were computed automatically by sponsored programs' spreadsheet.