Using Magpie for request extraction and workload modelling
Rebecca Isaacs
Tools to understand complex system behaviour are essential for many
performance analysis and debugging tasks. We have developed one such
tool, called Magpie, which automatically extracts a system's workload
under realistic operating conditions by tracking the control flow and
resource consumption of each and every request. From canonical
descriptions of these requests in which scheduling artefacts have been
removed, we can construct concise workload models suitable for
performance prediction and change detection. In this talk I will
describe Magpie's event correlation technique and present some
evaluation of its ability to accurately extract requests and to
construct representative models of system behaviour.
|