Efficient Database Query Processing on Modern Hardware
Website University of New Brunswick
About the Project
Project description:
The rapid advances in hardware technology is changing the landscape of database management and big data processing. With the rising main memory capacity, in-memory query processing is expected to become mainstream in commercial DBMSs. Conventional database execution engines evaluate a database SQL query by interpreting the query execution plan generated by a query planner. However, since each operator is implemented statically, it needs to be very generic in order to support all possible queries. Such generality may introduce significant runtime inefficiency, especially in the context of in-memory data management.
In this project, the PhD student will develop a query compiler for a commercial database system. The main objective is to execute queries using Just-In-Time (JIT) compilation in a manner that utilizes modern compilation techniques, and hardware features such as vectorization and prefetching. This is a fully funded (i.e. full scholarship) PhD position.
Location:
The University of New Brunswick, Fredericton is one of the top comprehensive universities of Canada. The Faculty of Computer Science is the first faculty of computer science in Canada and a leader in Atlantic Canada since 1968 with the oldest and most successful COOP program in Atlantic Canada.
Technical skills:
This research project will develop efficient database query execution engine using compiler and runtime techniques. The researcher will explore high performance SQL query processing approaches using cutting-edge query compilation techniques. Therefore, good knowledge in database internals and compiler design is desired. Systems programming skills, including programming in C/C++ in a Linux environment is required. Familiarity with LLVM Compiler infrastructure and Linux systems programming is quite beneficial.
Academic qualifications:
A solid background in Computer Science (or Computer Engineering), including a thesis-based research Master’s level degree from a reputed university with excellent grades, is required. Exceptional students with a non-traditional background having the required technical skills may be given consideration. Previous industrial/work experience in database systems internals will be considered an asset.
Contact:
Please contact with your CV, and Bachelor’s and Master’s degree transcripts; Email to [email protected]
To apply for this job email your details to bigdata@unb.ca