We demonstrate PerfEnforce, a dynamic scaling engine for analytics services. PerfEnforce automatically scales a cluster of virtual machines in order to minimize costs while probabilistically meeting the query runtime guarantees offered by a performance-oriented service level agreement (SLA). The demonstration will show three families of dynamic scaling algorithms --feedback control, reinforcement learning, and online machine learning--and will enable attendees to change tuning parameters, performance thresholds, and workloads to compare and contrast the algorithms in different settings.

There are known benchmarks for the performance evaluation of relational and XML databases. However, there is an increasing demand for database applications that require access to heterogeneous loosely-coupled distributed data sources . This paper presents a hybrid benchmark based on TPC-H where the data sources are heterogeneous. Specifically, the paper describes the design of the relational and XML data sources as well as the query redesign in the LINQ query language , which supports queries over heterogeneous data sources. The results of a performance evaluation of the hybrid benchmark over various database products is included for untyped and typed XML with and without clearing the database cache .

In this demonstration, we will showcase Myria, our novel cloud service for big data management and analytics designed to improve productivity. Myria's goal is for users to simply upload their data and for the system to help them be self-sufficient data science experts on their data -- self-serve analytics. Using a web browser, Myria users can upload data, author efficient queries to process and explore the data, and debug correctness and performance issues. Myria queries are executed on a scalable, parallel cluster that uses both state-of-the-art and novel methods for distributed query processing. Our interactive demonstration will guide visitors through an exploration of several key Myria features by interfacing with the live system to analyze big datasets over the web.