Princeton
University Office of Information Technology
Position Statement on Buy, Build, or
Open Source Software Decisions
A recent article in Syllabus Magazine
(November, 2003, The FREE, 0% APR,
Better Sex, No Effort Diet by Howard Strauss) has generated quite
a firestorm and touched a raw nerve in the open source community. The
author has been a long-time employee of Princeton, serving as manager
of user support, manager of advanced applications, and currently as
manager, technology outreach. He has written broadly on a number of
topics, some of which have been controversial and do not necessarily
represent the policy or technical directions of the Office of
Information Technology. The views expressed in the Syllabus article
were that of the author and not those of Princeton University nor its
Office of Information Technology. While the article might have led some
to believe that Princeton University opposes and does not support the
open source movement, in fact Princeton is an active participant in
open source activities, while continuing to run many university
services on vendor-supported commercial software.
Open Source Initiatives:
• Princeton was a founding member of the Java
Applications Special Interest Group (JA-SIG) and participant in the
uPortal open source effort to develop portals for higher education.
• Princeton faculty and staff have contributed to a
number of open source projects initiatives, including many cooperative
efforts that predated the current open source movement. Collaboration
among academic and research professionals has been vital to the
development of the current computing environment.
• Princeton University, along with 24 other major
research universities, have contributed to an effort, with funding from
the Mellon Foundation and the Open Software Application Foundation, to
develop an open source integrated e-mail, calendaring, and personal
information manager (PIM) application for higher education.
• Open source software is widely used at Princeton,
including Apache web server software (including the main
www.princeton.edu web site), Linux servers for infrastructure services
and desktop productivity, sendmail and spam assassin for mail
processing, and a number of Beowulf (clustered Linux) servers for high
performance computing.
Commercial Software:
• Most of Princeton’s infrastructure and
administrative computing runs on commercial software:
o Solaris and Windows servers provide most of the
main computing services
o Peoplesoft financial and human resources software
running on Oracle database provide the majority of administrative
services
o SunOne Directory and Messaging (e-mail) and
Microsoft Active Directory and Exchange provide authentication and
electronic mail for the majority of the campus
• Microsoft Windows and Macintosh operating systems
are used on most administrative, student, and faculty computer systems.
• Solaris Unix and other commercial Unix systems are
also in use for academic desktop along side a growing number of Linux
systems.
• Departmental servers run Windows, OS X, Solaris,
and other versions of Unix, including with Linux.
Locally Developed Software:
• A small number of administrative and support
systems have been locally developed (e.g., undergraduate admissions
application, loans and receivables, and helpdesk trouble-ticket
tracking systems).
Buy, Build, or Open Source
Decisions to run open source or commercial software should be based on
a number of criteria including availability, features, true costs,
mission criticality, and support. It should be remembered that open
source may not always be the least expensive, nor will commercial
products always provide the best support. When neither vendor nor open
source can meet the need, applications may need to be developed
locally.
When an organization is making a software decision, it must assess the
viability of vendor solutions, momentum of open source options, and
potential need to develop locally. None of the alternatives is
inherently “better” than the others. Each needs to be assessed relative
to the problem to be solved, appropriateness, and relevant internal and
external skills available.
Daniel J. Oberst (oberst@princeton.edu)
Director, Enterprise Infrastructure Services
Office of Information Technology
Princeton University
November 14, 2003