FMI: Fast and Cheap Message Passing for Serverless Functions
Open Access
- 21 June 2023
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 373-385
Abstract
Serverless functions provide elastic scaling and a fine-grained billing model, making Function-as-a-Service (FaaS) an attractive programming model. However, for distributed jobs that benefit from large-scale and dynamic parallelism, the lack of fast and cheap communication is a major limitation. Individual functions cannot communicate directly, group operations do not exist, and users resort to manual implementations of storage-based communication. This results in communication times multiple orders of magnitude slower than those found in HPC systems. We overcome this limitation and present the FaaS Message Interface (FMI). FMI is an easy-to-use, high-performance framework for general-purpose point-to-point and collective communication in FaaS applications. We support different communication channels and offer a model-driven channel selection according to performance and cost expectations. We model the interface after MPI and show that message passing can be integrated into serverless applications with minor changes, providing portable communication closer to that offered by high-performance systems. In our experiments, FMI can speed up communication for a distributed machine learning FaaS application by up to 162x, while simultaneously reducing cost by up to 397 times.Keywords
All Related Versions
Funding Information
- European Research Council (101002047)
- Horizon 2020 Framework Programme (955776, 955606)
This publication has 42 references indexed in Scilit:
- Searching for exotic particles in high-energy physics with deep learningNature Communications, 2014
- Two-tree algorithms for full bandwidth broadcast, reduction and scanParallel Computing, 2009
- Performance analysis and optimization of MPI collective operations on multi-core clustersThe Journal of Supercomputing, 2009
- Bandwidth optimal all-reduce algorithms for clusters of workstationsJournal of Parallel and Distributed Computing, 2009
- MPI collective algorithm selection and quadtree encodingParallel Computing, 2007
- Collective communication: theory, practice, and experienceConcurrency and Computation: Practice and Experience, 2007
- Performance analysis of MPI collective operationsCluster Computing, 2007
- Parallel Prefix (Scan) Algorithms for MPI2006
- Send-receive considered harmfulACM Transactions on Programming Languages and Systems, 2004
- Scans as primitive parallel operationsIEEE Transactions on Computers, 1989