Custom Serverless Function Scheduling Policies: An APP Tutorial

State-of-the-art serverless platforms use hard-coded scheduling policies that hardly accommodate users in implementing functional or performance-related scheduling logic of their functions, e.g., preserving the execution of critical functions within some geographical boundaries or minimising data-access latencies. We addressed this problem by introducing APP: a declarative language for defining per-function scheduling policies which we also implemented as an extension of the open-source OpenWhisk serverless platform. Here, we present a gentle introduction to APP through an illustrative application developed over several incremental steps.