abstract
- Notebooks Data processing systems are increasingly being deployed in the cloud, because of the benefits of elasiticity and short-term resource provisioning. In recent years, serverless cloud computing is offered in the form of highly elastic resource pools. However, analyzing and understanding the performance and cost characteristics of serverless cloud infrastructure in the context of data processing is challenging. In this demonstration, we present our Skyrise evaluation framework for experimentation in serverless data processing. The framework provides a suite of microbenchmarks and a serverless query engine to run end-to-end workloads. Users can benchmark serverless compute and storage resources and observe the impact of their performance on dataintensive applications. In addition, they can study the trade-offs compared to server-based systems. Utilizing the inherent elasticity of serverless resources, our framework facilitates interactive analysis and supports the understanding of key characteristics of serverless data processing for data system engineers. 1 Results Experiment Driver Tasks Network I/O Function Query Worker Application Level Parameters Storage I/O Function FaaS/IaaS Platform Minimal Function Resource Level Figure 1: Architecture of the Skyrise evaluation framework showing the experiment execution flow from config to plot. TPC-H and TPCx-BB benchmarks. Query result metrics comprise the wall-clock and the cumulative compute times, and the cost of any used cloud resources. During this demonstration, we let users (1) run experiments showing characteristic features of serverless cloud infrastructure and (2) observe how these translate to application performance. The results are summarized in Jupyter notebooks to (3) aid in comprehension and comparison with server-based applications. We invite users to explore (4) different Skyrise benchmarks and benchmark configurations. We summarize our demonstration as follows: