WiseFuse: Workload Characterization and DAG Transformation for ServerlessWorkflows

We characterize production workloads of serverless DAGs at a major cloud provider. Our analysis highlights two major factors that limit performance: (a) lack of efficient communication methods between the serverless functions in the DAG, and (b) stragglers when a DAG stage invokes a set of parallel functions that must complete before starting the next DAG stage. To address these limitations, we propose WiseFuse, an automated approach to generate an optimized execution plan for serverless DAGs for a user-specified latency objective (SLO) or cost budget. We introduce three optimizations: (1) Fusion combines in-series functions together in a single VM to reduce the communication overhead between cascaded functions. (2) Bundling executes a group of parallel invocations of a function in one VM to improve resource sharing among the parallel invocations to reduce skew. (3) Resource Allocation assigns the right size to each VM hosting a function or a group of functions to reduce the latency and cost of invoking the serverless DAG. We implement WiseFuse and evaluate it experimentally using three serverless applications, namely, Video Analytics, Approximate SVD, and ML Analytics, which span different DAG structures, memory footprints, and intermediate data sizes. In comparison to competing approaches, WiseFuse shows significant improvements in E2E latency and cost. Specifically, for the ML pipeline, WiseFuse achieves a P95 latency that is 67% lower than Photons [SoCC-20], 39% lower than Faastlane [USENIX ATC-21], and 90% lower than Sonic [USENIX ATC-21], without increasing the $ cost.