Event-Driven Machine Learning Infrastructure: Performance Benchmarking of AWS Lambda and Fargate Serverless Compute
Keywords:
AWS Lambda; AWS Fargate; Serverless Computing; Event-Driven Architecture; Machine Learning Inference; Performance Benchmarking; Cold Start; Execution Latency; Throughput; Cost Analysis.Abstract
This paper assessed AWS Fargate and AWS Lambda as serverless compute platforms for running event-driven machine learning inference tasks. To mimic real-time event processing scenarios, both platforms were benchmarked under the same settings using a common ML model and a variety of input payload sizes. Measured and examined key performance indicators—including cold start delay, execution time, throughput, and cost-efficiency. The findings showed that AWS Lambda had quicker execution times for smaller payloads and better scalability under high concurrency, whereas AWS Fargate had shorter cold start latency across all resource configurations. While AWS Fargate grew more affordable for bigger, long-running jobs, cost study showed AWS Lambda was more affordable for lightweight, short-duration operations. The results underlined the need of choosing compute platforms depending on particular workload needs since they showed important trade-offs between performance and cost. This benchmarking study offers valuable insights for architects and developers designing scalable, event-driven ML systems in cloud-native environments.
Downloads
References
A. Rose, Performance Evaluation of Serverless Object, Ph.D. dissertation, California State University, Northridge, 2023.
B. N. Y. Arafath, A comparative study between microservices and serverless in the cloud, Master's thesis, OsloMet-storbyuniversitetet, 2022.
C. Lekkala, "Containerization vs. Serverless Architectures for Data Pipelines," Serverless Architectures for Data Pipelines, Feb. 1, 2023.
D. M. Naranjo, S. Risco, G. Moltó, and I. Blanquer, "A serverless gateway for event‐driven machine learning inference in multiple clouds," Concurrency and Computation: Practice and Experience, vol. 35, no. 18, p. e6728, 2023.
G. Kambala, "Cloud-Native Architectures: A Comparative Analysis of Kubernetes and Serverless Computing," 2023.
I. Goswami, "Serverless Architecture for Machine Learning," 2023.
J. J. Paul, Distributed Serverless Architectures on AWS, Berkeley, CA, 2023.
L. Scotton, Engineering framework for scalable machine learning operations, 2021.
M. Rahman, "Serverless cloud computing: a comparative analysis of performance, cost, and developer experiences in container-level services," 2023.
M. Sisák, Cost-optimal AWS Deployment Configuration for Containerized Event-driven Systems, Ph.D. dissertation, 2021.
N. Kodakandla, "Serverless Architectures: A Comparative Study of Performance, Scalability, and Cost in Cloud-native Applications," Iconic Research and Engineering Journals, vol. 5, no. 2, pp. 136-150, 2021.
P. Grzesik, D. R. Augustyn, Ł. Wyciślik, and D. Mrozek, "Serverless computing in omics data analysis and integration," Briefings in Bioinformatics, vol. 23, no. 1, p. bbab349, 2022.
S. Eismann, Performance Engineering of Serverless Applications and Platforms, Ph.D. dissertation, Universität Würzburg, 2023.
S. R. Gallardo, Serverless strategies and tools in the cloud computing continuum, Ph.D. dissertation, Universitat Politècnica de València, 2023.
V. Naik, Machine Learning Using Serverless Computing, 2021.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.
