The Topology-aware Allocation Priority Policies (tAPP) language allows users of serverless platforms to orient the scheduling of their functions w.r.t. the topological properties of the available computation nodes. A tAPP-based platform can support multiple scheduling policies, which one would usually enforce via (brittle) ad-hoc multi-instance platform deployments. In this paper, we present an extension of the Apache OpenWhisk serverless platform that supports tAPP-based scripts. We show that our extension does not negatively impact the performance of generic, non-topology-bound serverless scenarios, while it increases the performance of topology-bound ones.
De Palma, G., Giallorenzo, S., Mauro, J., Trentin, M., Zavattaro, G. (2024). An OpenWhisk Extension for Topology-Aware Allocation Priority Policies. GEWERBESTRASSE 11, CHAM, CH-6330, SWITZERLAND : Springer Science and Business Media Deutschland GmbH [10.1007/978-3-031-62697-5_11].
An OpenWhisk Extension for Topology-Aware Allocation Priority Policies
De Palma, Giuseppe;Giallorenzo, Saverio;Trentin, Matteo;Zavattaro, Gianluigi
2024
Abstract
The Topology-aware Allocation Priority Policies (tAPP) language allows users of serverless platforms to orient the scheduling of their functions w.r.t. the topological properties of the available computation nodes. A tAPP-based platform can support multiple scheduling policies, which one would usually enforce via (brittle) ad-hoc multi-instance platform deployments. In this paper, we present an extension of the Apache OpenWhisk serverless platform that supports tAPP-based scripts. We show that our extension does not negatively impact the performance of generic, non-topology-bound serverless scenarios, while it increases the performance of topology-bound ones.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.