Systar is a provider of performance management softwares that enable large organizations to maximize the efficiency of their business operations and IT infrastructure.
Systar is currently developping a new product incorporating a distributed, replicated and column oriented database system for bitemporal data. This system faces several functional and performance constraints, e.g managing important volume of data and processing a high volume of concurrent queries.
The thesis objective is to optimize Systar's product. It consists in finding a good architecture, good data structures and algorithms to efficiently process high data volume(hundred of thousands of events per second). The most interesting propositions will be implemented and approved thanks to a testbed used by Systar.