Tutorials

Benchmark Big Data Analytics Systems

April 20, 2020, Morning Session

Authors and Affiliations:

  • Rekha Singhal1 (rekha.singhal@tcs.com)
  • Todor Ivanov2 (todor@dbis.cs.uni-frankfurt.de)
    1Tata Consultancy Services, India and 2Goethe University Frankfurt, Germany

Summary:

There is need to understand how to benchmark systems used to build AI based solutions. AI based solutions have a complex pipeline of pre-processing, statistical analysis, machine learning and deep learning on data to build prediction models. The performance metrics may be data pre- processing time, model training time, model inference time or model accuracy. We do not see a single benchmark answering all questions of solution architects and researchers. This tutorial covers both practical and research questions on relevant Big Data and Analytics benchmarks.

Performance Engineering for Microservices and Serverless Applications: the RADON approach

April 20, 2020, Afternoon Session

Authors and Affiliations:

  • A. U. Gias1 (a.gias17@imperial.ac.uk)
  • A. van Hoorn2 (van.hoorn@informatik.uni-stuttgart.de)
  • L. Zhu1 (lulai.zhu15@imperial.ac.uk)
  • G. Casale1 (g.casale@imperial.ac.uk)
  • T. F. Düllmann2 (duellmann@iste.uni-stuttgart.de)
  • M. Wurster2 (michael.wurster@iaas.uni-stuttgart.de)
    1Imperial College London, UK and 2University of Stuttgart, Germany

Summary:

This tutorial presents the performance engineering approach for microservices and serverless applications being developed as part of the RADON project (http://radon-h2020.eu). First, we will introduce microservices and serverless FaaS applications, and performance engineering challenges in this context. The focus will then shift to how these applications can be modeled using TOSCA and how different performance specifications can be integrated into such models. After that, the tutorial will discuss how such models are utilized in RADON to determine optimal decomposition strategies of serverless functions. Once the application gets deployed, following the DevOps practice, it is necessary to generate and maintain performance test cases for continuous integration and deployment. Finally, the tutorial focuses on runtime resource management by demonstrating a fine-grained autoscaling approach.