Note: [] Address of correspondence: Nong Ye, Professor of Industrial
Engineering, Arizona State University, Information and Systems Assurance
Laboratory, Box 875906, Tempe, Arizona 85287-5906, USA. Tel.: +1 480 965 7812;
Fax: +1 480 965 8692; E-mail: nongye@asu.edu
Abstract: An increasing number of business transactions, mission-critical
operations and many other activities rely on the services of computer and
network systems, and the dependability of those services. However, many
computer and network systems do not currently provide adequate service
dependability. Such systems place no admission control on service requests
(jobs), leading to variability and thus unpredictability in job waiting times
which have a negative impact on service dependability. This paper presents a
Batch Scheduled Admission Control (BSAC) method to help provide predictability
in job waiting time for high priority jobs. With this method, we change the
dynamic job arriving problem into a static problem which allows the
introduction of job scheduling techniques that require a static set of jobs to
operate on for stabilizing their waiting time. We illustrate how our method can
be employed through examples and experiments that compare the performance of
our method with that of no admission control. The results under our test cases
show that our method improves the variance of job waiting times, and in some
cases with little sacrifice to the mean of job waiting times.
Keywords: Admission control, quality of service, computer networks