ICSA Colloquium Talk - 26/10/2021
Talk Title
Turbocharging Serverless Research with STeLLAR and vHive
Abstract Serverless has emerged as the next dominant cloud architecture and programming paradigm thanks to its high scalability and flexible, pay-as-you-go billing model. In serverless, developers structure their cloud services as a collection of functions while providers take responsibility for scaling each function’s resources according to traffic changes. This labor division opens up great opportunities for systems researchers who seek to innovate in serverless computing. Unfortunately, leading serverless providers, like AWS Lambda and Azure Functions, continue to rely on proprietary infrastructure, leaving academics to black-box research with production offerings or building their in-house prototypes of various serverless infrastructure components. Our work bridges this gap between the industry and academia, equipping systems researchers with representative open-source frameworks for performance analysis and cross-stack innovation in serverless. The first one is STeLLAR, a provider-agnostic framework for benchmarking production clouds, which unlocks detailed performance analysis of proprietary commercial clouds. Using STeLLAR to study three leading cloud platforms, we identified that modern serverless infrastructure’s performance is dominated by data movement, which is inherent to serverless function cold starts and cross-function communication. To evaluate and prototype state-of-the-art serverless systems, we built vHive, a full-stack open-source framework and ecosystem for serverless experimentation and innovation. vHive is representative of real serverless clouds, as recognized by two leading serverless providers and an ASPLOS'21 Distinguished Artifact Award, integrating production-grade components.
Bio-s: Dmitrii is a final-year Ph.D. student at the University of Edinburgh, co-advised by Prof. Boris Grot and Prof. Edouard Bugnion (EPFL). Dmitrii’s research interests lie at the intersection of Computer Systems and Architecture with a current focus on support for cloud and serverless computing. Theodor received his B.Sc. degree in AI & CS from the University of Edinburgh, completing his thesis under the supervision of Prof. Boris Grot and Dmitrii Ustiugov. He is now a first-year M.Sc. student at ETH Zurich, currently working at the intersection of Computer Systems and Machine Learning.
ICSA Colloquium Talk - 26/10/2021
Hybrid (G.03, IF or via Zoom link)