|Researchers at Purdue University have developed ORION, a multifaceted approach to optimize execution paths and reduce latency for serverless cloud workflows. This is achieved through the combined use of a distribution and correlation aware performance model for end-to-end (E2E) latency, resource optimization, pre-warming strategy, and bundling multiple invocations of a function. In comparison with alternate approaches (Photons, Faastlane, Sonic), ORION offers a latency improvement of up to 90% without increasing financial cost. When targeted for cost reduction, ORION achieves up to 53% reductions.
- Combines multiple optimization techniques
- Reduced E2E latency (39-90% faster than alternatives)
- Financial cost reductions
- Lower computational cost
- Cloud computing
- Serverless cloud applications
- Big data
- Microsoft Azure, AWS Lambda
This technology has been validated through testing of ORION and comparing it to other approaches.
- Ashraf Mahgoub, Edgardo Barsallo Yi, Karthick Shankar, Sameh Elnikety, Somali Chaterji, and Saurabh Bagchi. 2022. ORION and the Three Rights: Sizing, Bundling, and Prewarming for Serverless DAGs. 16th USENIX Symposium on Operating Systems Design and Implementation (OSDI 22), Carlsbad, CA.
- Ashraf Mahgoub, Edgardo Barsallo Yi, Karthick Shankar, Eshaan Minocha, Sameh Elnikety, Saurabh Bagchi, and Somali Chaterji. 2022. WiseFuse: Workload Characterization and DAG Transformation for Serverless Workflows. In Abstract Proceedings of the 2022 ACM SIGMETRICS/IFIP PERFORMANCE Joint International Conference on Measurement and Modeling of Computer Systems (SIGMETRICS/PERFORMANCE '22 Abstracts), June 6–10, 2022, Mumbai, India. ACM, Mumbai, India, 2 pages. https://doi.org/10.1145/3489048.3530959
May 26, 2022
Purdue Office of Technology Commercialization
The Convergence Center
101 Foundry Drive, Suite 2500
West Lafayette, IN 47906
Phone: (765) 588-3475
Fax: (765) 463-3486