Systems Implementation of the Integrated Public Use Microdata Series:  
A Description of the Computing Infrastructure of the IPUMS

This presentation will describe the computing infrastructure of the
Integrated Public Use Microdata Series (IPUMS).  The IPUMS is a coherent
series of individual-level U.S.  census data drawn from 13 federal
censuses between 1850 and 1990 and is used by social science researchers
worldwide (for more information on IPUMS, see www.ipums.umn.edu).  The
computers that currently serve IPUMS to the world are a mix of Sun and
Intel architecture running various Unix operating systems, Fortran and
Perl programs, shell scripts, and HTML.  In spite of its apparent
complexity, however, IPUMS development has been responsive to user input,
packaged into a format suitable for installation at mirror sites, and
successful at maximizing its computing capability at minimal cost and
software-licensing overhead.

We will begin by describing the IPUMS Data Extraction System, which has
been available on the World Wide Web for over three years and undergoes
continuous development.  User suggestions have led to a number of
improvements in the system, such as the ability for users to retrieve and
modify old extract requests.  The amount of online documentation available
to users of the IPUMS Data Extraction System has also greatly increased.  
This paper will describe these and other recent software enhancements to
the IPUMS Data Extraction System. Next we will introduce the software and
features of IPUMS that comprise its ability to function in a compute
cluster mode.  The cluster design grew out of a need for greater computing
power with minimal resources and has been in successful operation for over
a year.  Additionally, after IPUMS started receiving inquiries about
establishing mirror sites of its Data Extraction System, we worked to make
IPUMS portable and installable across multiple platforms.  The IPUMS job
scheduling feature will also be explained, as well as the process by which
IPUMS is packaged for distribution and installation at remote sites.

Last, we will introduce the latest advances in IPUMS computing and outline
plans for future development of the system.  This includes demonstrating a
full-fledged stand- alone installation of IPUMS on a laptop that boots
Linux and runs a virtual Windows operating system as a task.  While
originally designed for demonstrating IPUMS at academic conferences, some
researchers would undoubtedly find such a setup useful.  Future
development plans for IPUMS include plans to re-implement the IPUMS Data
Extraction System using Java and the incorporation of DTD and
XML-compliant metadata coding schemes throughout the IPUMS.

William C. Block
25 Blegen Hall
269 19th Avenue South
Minneapolis, MN 55455
(612) 624-7091 Voice
(612) 625-4814 Fax
block@socsci.umn.edu

John T. Easton
25 Blegen Hall
269 19th Avenue South
Minneapolis, MN  55455
(612) 625-1079 Voice
(612) 625-4814 Fax
easton@socsci.umn.edu

Todd K. Gardner
614 Social Science Tower
267 19th Avenue South
Minneapolis, MN 55455
(612) 624-5818
(612) 624-7096
gardner@socsci.umn.edu