Allocation Priority Policies for Serverless Function-Execution Scheduling Optimisation

Giuseppe de Palma (1) , Saverio Giallorenzo (1, 2) , Jacopo Mauro (3) , Gianluigi Zavattaro (1, 2)

1 UNIBO - Alma Mater Studiorum Università di Bologna = University of Bologna
2 FOCUS - Foundations of Component-based Ubiquitous Systems
3 SDU - University of Southern Denmark

Serverless computing is a Cloud development paradigm where developers write and compose stateless functions, abstracting from their deployment and scaling. In this paper, we address the problem of functionexecution scheduling, i.e., how to schedule the execution of Serverless functions to optimise their performance against some user-defined goals. We introduce a declarative language of Allocation Priority Policies (APP) to specify policies that inform the scheduling of function execution. We present a prototypical implementation of APP as an extension of Apache OpenWhisk and we validate it by i) implementing a use case combining IoT, Edge, and Cloud Computing and ii) by comparing its performance to an alternative implementation that uses vanilla OpenWhisk.

Mots clés: Serverless; Function-execution Scheduling; Optimisation

Domaines: Langage de programmation [cs.PL]

Giuseppe de Palma, Saverio Giallorenzo, Jacopo Mauro, Gianluigi Zavattaro. Allocation Priority Policies for Serverless Function-Execution Scheduling Optimisation. ICSOC 2020 - 18 th International Conference on Service-Oriented Computing, Dec 2020, Dubai, United Arab Emirates. pp.416 - 430, ⟨10.1007/978-3-030-65310-1_29⟩. ⟨hal-03347275⟩