MoN18: Eighteenth Mathematics of Networks meeting

Peter Thompson (Predictable Network Solutions) A novel approach to verification of distributed system performance

A key risk for time-critical distributed systems is that the performance will be unacceptable and/or uneconomic to deliver. Mitigating this risk requires an approach that supports both prospective validation and post-facto verification that should: capture performance requirements and resource constraints; validate them; construct performance properties/invariants that witness those requirements; and support the reification and abstraction of such properties/invariants throughout the System Development Life-Cycle. Here, we present a novel approach to prospective performance and resource consumption validation called the ∆Q framework, based on the use of improper random variables, which addresses these issues.

Return to previous page

Contact: Keith Briggs (mailto:keith.briggs_at_bt_dot_com) or Richard G. Clegg (