|
In 2003, with funding from a Federal Earmark grant and the Dean
of the Charles E. Schmidt College of Science, the Mathematics
Department configured, acquired and put into operation a 25 node
computer cluster located in the “cryptology lab” SE 2XX. The goal has
been to use the cluster for graduate, undergraduate and faculty
research and teaching in cryptology, coding theory, combinatorics, and
related areas. However, the cluster’s utility/functionality has been
rapidly deteriorating over the past 6 years, as the hardware became
increasingly obsolete. The purpose of our present request is to
refurbish the aging cluster by replacing the 1GHz/1GB nodes by
present-day multicore 3GHz/16GB processors which will provide much higher
performance at a fraction of the original cost. UPS power supplies
and the interconnecting switching gear will not need replaced,
although some funds are needed to refurbish the UPS power supplies.
In its early days the cluster was used mostly in connection with
the development of parallel algorithms, particularly needed in the
solution of discrete mathematical problems, considered computationally
intractable. Such problems are “lattice basis reduction”, “group
factorizations”, “integer and polynomial factorization”, the “discrete
logarithm problem”, “DNA sorting permutations by transpositions in
genome rearrangements”, and finding “collisions in cryptographic hash
functions”. Presently, the cluster is as sorely inadequate for
research as is for teaching. The refurbished and upgraded cluster will
be used for research by students at the Ph.D. and M.Sc. level, by
faculty, and also for undergraduate research as it relates to QEP.
While the availability of a working cluster is imperative for the
research of at least 3 of our Ph.D. students, the cluster will also be
used to introduce students to parallel computation solutions and the
development of parallel algorithms in CIS 4362 (Cryptography and
Information Security), MAD 4301 (Graph Theory),MAD 4605 (Introduction
to Coding Theory), STA 4618 (Linear Programming and Game Theory), MAD
5474 (Intro to Cryptology and Information Security), MAD 5202
(Introductory Combinatorics), MAT 6715 (Mathematics and Technology),
MAD 6209 (Topics in Combinatorics), MAD 6307 (Graph Theory), MAD 6477
(Cryptography), MAD 6478 (Cryptanalysis), MAD 6607 (Coding Theory).
Some introductory use is also desirable in MAD 2502 (Computational
Mathematics). Happily, we have a partnership with the software
companies “APL2000/APL Next” who are providing us the needed software
at no cost, under educational license, and will provide us with the
latest “parallel manager software” for the cluster. Thus, we will not
incur additional software costs. |
