Self-optimizing Systems

Tuneful

Tuneful is an extension for Spark which optimizes workload configurations starting from a zero-knowledge setting. The more workloads a cluster executes, the better it becomes at executing them. In order to achieve this, we leverage Multi Task Gaussian Process, Similarity Analysis and Significance Analysis.

Papers
  1. To Tune or Not to Tune? In Search of Optimal Configurations for Data Analytics from ACM KDD 2020 (link)
  2. Accelerating the Configuration Tuning of Big Data Analytics with Similarity-aware Multitask Bayesian Optimization from IEEE BigData 2020 (link)
  3. Towards Seamless Configuration Tuning of Big Data Analytics from IEEE ICDCS 2019 (link)
Code
  1. Tuneful source code
  2. Simtune source code
Data
  1. KDD dataset
People
  1. Thomas Pasquier
arrow_back Back

Systopia lab is supported by a number of government and industrial sources, including Cisco Systems, the Communications Security Establishment Canada, Intel Research, the National Sciences and Engineering Research Council of Canada (NSERC), Network Appliance, Office of the Privacy Commissioner of Canada, and the National Science Foundation (NSF).