Dr. Wei Wang and PhD student Suyi Li Won the Best Paper Award from ACM SoCC 2023
CSE PhD student Suyi Li, in collaboration with his advisor Dr. Wei Wang and other researchers from WeBank, has achieved a significant milestone. Their paper, entitled "Golgi: Performance-Aware, Resource-Efficient Function Scheduling for Serverless Computing," received the Best Paper Award at the 2023 ACM Symposium on Cloud Computing (SoCC), a leading research conference on cloud systems. This accolade holds special significance as it marks the first time a Hong Kong-based university has been recognized by this prestigious award.
About the Award-winning Paper
This award-winning paper introduces Golgi, a novel scheduling system designed for serverless functions, with the goal of minimizing resource provisioning costs while meeting the function latency requirements. To achieve this, Golgi judiciously over-commits functions based on their past resource usage. To ensure overcommitment does not cause significant performance degradation, Golgi identifies nine low-level metrics to capture the runtime performance of functions, encompassing factors like request load, resource allocation, and contention on shared resources. These metrics enable accurate prediction of function performance using the Mondrian Forest, a classification model that is continuously updated in real-time for optimal accuracy without extensive offline training. Golgi employs a conservative exploration-exploitation strategy for request routing. By default, it routes requests to non-overcommitted instances to ensure satisfactory performance. However, it actively explores opportunities for using more resource-efficient overcommitted instances, while maintaining the specified latency SLOs (Service-Level Objectives). Golgi also performs vertical scaling to dynamically adjust the concurrency of overcommitted instances, maximizing request throughput and enhancing system robustness to prediction errors.
The researchers have prototyped Golgi and evaluated it in both AWS EC2 cluster and a small production cluster. The results show that Golgi can meet the SLOs while reducing the resource provisioning cost by 42% (30%) in EC2 cluster (their production cluster).
About ACM SoCC
SoCC'23 is 14th edition of the annual ACM Symposium on Cloud Computing, the premier conference on cloud computing. It brings together researchers, developers, users, and practitioners interested in cloud computing, and is the only conference co-sponsored by the ACM Special Interest Groups on Management of Data (SIGMOD) and on Operating Systems (SIGOPS). As in prior SoCC conferences, SoCC'23 will foster connections between the academic and industrial communities operating in the cloud space.
Congratulation once again to Dr. Wang, Mr. Suyi Li, and their fellow co-authors on this outstanding achievement.