Talk: Performance Challenges of FaaS Cloud Architectures @ ICPE 2018

April 12, 2018 · 1 minute read

As part of the serverless effort of the SPEC Cloud Research Group we have been working towards a benchmark of serverless frameworks. We published an initial vision for the field of serverless computing, which includes a consistent terminology and high-level perspectives and challenges for the field. Increasingly focussing our scope, we decided to dive deeper into the challenges that fit the domain of the research group: performance-related challenges. This lead to the submission (and acceptance!) of our second vision paper: A SPEC RG Cloud Group’s Vision on the Performance Challenges of FaaS Cloud Architectures

I presented the paper at ICPE 2018 in a short presentation (7 min time slot) in main session. With the constraint time slot, the presentation focused on three key aspects of the paper: (1) introduce and provide motivation for this new ‘serverless’ and ‘FaaS’ model; (2) provide a brief overview of the performance challenges that we found; (3) and, discuss the current and future work of the SPEC RG CLOUD serverless activity.

Abstract As a key part of the serverless computing paradigm, Function-as-a-Service (FaaS) platforms enable users to run arbitrary functions without being concerned about operational issues. However, there are several performance-related issues surrounding the state-of-the-art FaaS platforms that can deter widespread adoption of FaaS, including sizeable overheads, unreliable performance, and new forms of the cost-performance trade-off. In this work we, the SPEC RG Cloud Group, identify six performance-related challenges that arise specifically in this FaaS model, and present our roadmap to tackle these problems in the near future. This paper aims at motivating the community to solve these challenges together.

Paper Slides