Abstract—Periodic processing of software components is a
simple reflection of a periodic nature of many real world
processes where the identical actions are repeated at every given
period of time. Discovering periodic patterns in the traces of
computations performed in the past allows for better
preparation of software systems to meet the demands of high
workload times in the future.
This work shows how to discover the periodic patterns in the
nested logs of computations using the transformations of the
simple patterns into the complex ones. The paper defines a
concept of periodic pattern and its validation in a reduced nested
log of events. A system of derivations rules is defined and
followed by a sequence of algorithms that use the rules to create
the complex periodic patterns. The paper is concluded with a
description of experiment where the sequences SQL statement
are transformed into the expressions of extended relational
algebra and used as input data to discover the periodic patterns
with a method described in the paper.
Index Terms—Periodic patterns, derivation rules, nested
events, nested logs.
J. R. Getta is with the School of Computing and Information Technology,
University of Wollongong, Wollongong, NSW 2530, Australia (e-mail:
jrg@uow.edu.au).
M. Zimniak is with Faculty of Computer Science, TU Chemnitz,
Chemnitz, Germany (e-mail: marcin.zimniak@cs.tu-chemnitz.de).
[PDF]
Cite: Janusz R. Getta and Marcin Zimniak, "Deriving Complex Periodic Patterns from Nested Logs of Events," International Journal of Knowledge Engineering vol. 1, no. 3, pp. 223-229, 2015.
