Owl: performance-aware scheduling for resource-efficient function-as-a-service cloud

This work documents our experience of improving the scheduler in Alibaba Function Compute, a public FaaS platform. It commences with our observation that memory and CPU are under-utilized in most FaaS sandboxes. A natural solution is to overcommit VM resources when allocating sandboxes, whereas the ensuing contention may cause performance degradation and compromise user experience. To complicate matters, the degradation in FaaS can arise from external factors, such as failed dependencies of user functions.

We design Owl to achieve both high utilization and performance stability. It introduces a customizable rule system for users to specify their toleration of degradation, and overcommits resources with a dual approach. (1) For less-invoked functions, it allocates resources to the sandboxes with usage-based heuristic, keeps monitoring their performance, and remedies any detected degradation. It differentiates whether a degraded sandbox is affected externally by separating a contention-free environment and migrating the affected sandbox into there as a comparison baseline. (2) For frequently-invoked functions, Owl profiles the interference patterns among collocated sandboxes and place the sandboxes under the guidance of profiles. The collocation profiling is designed to tackle the constraints that profiling has to be conducted in production. Owl further consolidates idle sandboxes to reduce resource waste. We prototype Owl in our production system and implement a representative benchmark suite to evaluate it. The results demonstrate that the prototype could reduce VM cost by 43.80% and effectively mitigate latency degradation, with negligible overhead incurred.