Virtually Timed Ambients for Cloud Computing
Johanna Beate Stumpf, Informatics
Cloud computing provides a virtual computing environment for executing tasks on pay-on-demand metered resources. Our research shows how performance and resource consumption on the cloud can be formally modeled and analysed. Our starting point is the ambient calculus, with its hierarchical locations for execution. We represent resource provisioning in nested ambients by an encoding of virtual time slices to reason about the timed behaviour of ambients. To compare single ambients and whole systems with respect to speed, we define weak timed bisimulation for mobile ambients as a conservative extension of untimed bisimulation and show that the equivalence of contextual bisimulation and reduction barbed congruence can be extended. The calculus for virtually timed mobile ambients is illustrated by examples.