PolyLARVA : runtime verification with configurable resource-aware monitoring boundaries

Abstract

Runtime verification techniques are increasingly being applied in industry as a lightweight formal approach to achieve added assurance of correctness at runtime. A key issue determining the adoption of these techniques is the overheads introduced by the runtime checks, affecting the performances of the monitored systems. Despite advancements in the development of optimisation techniques lowering these overheads, industrial settings such as online portals present new challenges, since they frequently involve the handling of high volume transaction throughputs and cannot afford substantial deterioration in the service they provide.

One approach to reduce overheads is the deployment of the verification computation on auxiliary computing resources, creating a boundary between the system and the verification code. This limits the use of system resources with resource intensive verification being carried out on the remote-side. However, under particular scenarios this approach may still not be ideal, as it may induce significant communication overheads. In this paper, we propose a framework which enables fine-tuning of the tradeoff between processing, memory and communication monitoring overheads, through the use of a user-configurable monitoring boundary. This approach has been implemented in the second generation of the Larva runtime verification tool, polyLarva.