Information for

Ashraf Aboulnaga

Adjunct Associate Professor

Research interestsAshraf Aboulnaga

Professor Aboulnaga's research is in the area of database management, with a current focus on databases in cloud computing environments, data integration on the web, and self-managing database systems.

The goal of Professor Aboulnaga's research in the area of cloud computing is to enable effective, scalable, and reliable database management in cloud computing environments. Part of this research focuses on virtualization, in particular tuning database systems running on virtual machines and using virtual storage, and also making database systems reliable and elastically scalable by taking advantage of the capabilities of virtual machine technologies. Another part of this research focuses on optimizing scalable analytics environments based on Map-Reduce style systems such as Hadoop by, for example, improving job scheduling or enabling the reuse of intermediate results.

Professor Aboulnaga's research in the area of data integration is motivated by the abundance of structured data on the web. It would be useful in many situations to provide a unified structured view (i.e., a unified schema) of the parts of this data that relate to a particular topic. However, at web scale it is not feasible to manually identify and integrate data sources about a given topic. Professor Aboulnaga's research therefore focuses on developing fully automatic and scalable techniques for data integration, with the key feature that these techniques start with a best-effort "good-enough" solution and refine this solution iteratively based on implicit or explicit feedback from the user. This style of data integration is sometimes referred to as pay-as-you-go data integration.

In the area of self-managing database systems, Professor Aboulnaga's recent research has focused on developing techniques that help database administrators choose good physical database designs. He has focused on automatic physical design for XML databases and on robust physical database designs for relational databases. His research has also addressed workload management issues such as modeling the interaction between concurrently running queries, optimally scheduling interacting queries, and managing database systems under overload.

Degrees and awards

BS, MS (Alexandria), MS, PhD (Wisconsin)

ACM Senior Member (2010); IEEE Senior Member (2010); Ontario Early Researcher Award (2009); Google Research Award (2008); IBM Centre for Advanced Studies (CAS) Project of the Year Award (2008); IBM CAS Faculty Fellow (2006-2011)

Industrial and sabbatical experience

Professor Aboulnaga spent six months in 2009-2010 as a Visiting Research Scientist at Google's Waterloo office, where he worked on optimizing the infrastructure for serving content advertisements.

From 2002 to 2004, Professor Aboulnaga was a Research Staff Member at the IBM Almaden Research Center in San Jose, California. He worked on the IBM DB2 LEarning Optimizer (LEO) project, addressing the problem of recommending database statistics based on feedback information from the execution of user queries, and transferring this technology into the IBM DB2 UDB V8.2 database management system.

Representative publications

Mumtaz Ahmad, Songyun Duan, Ashraf Aboulnaga, and Shivnath Babu. Predicting Completion Times of Batch Query Workloads using Interaction-aware Models and Simulation. Proceedings of the International Conference on Extending Database Technology (EDBT), 2011.

Sean Tozer, Tim Brecht, and Ashraf Aboulnaga. Q-Cop: Avoiding Bad Query Mixes to Minimize Client Timeouts Under Heavy Loads. Proceedings of the IEEE International Conference on Data Engineering (ICDE), 2010.

Hatem A. Mahmoud and Ashraf Aboulnaga. Schema Clustering and Retrieval for Multi-domain Pay-as-you-go Data Integration Systems. Proceedings of the ACM SIGMOD International Conference on Management of Data, 2010.

Ahmed A. Soror, Umar Farooq Minhas, Ashraf Aboulnaga, Kenneth Salem, Peter Kokosielis, and Sunil Kamath. Automatic Virtual Machine Configuration for Database Workloads. ACM Transactions on Database Systems (TODS), vol 35, no 1, 2010.

Xin Liu, Ashraf Aboulnaga, Kenneth Salem, and Xuhui Li. CLIC: CLient-Informed Caching for Storage Servers. Proceedings of the USENIX Conference on File and Storage Technologies (FAST), 2009.

Iman Elghandour, Ashraf Aboulnaga, Daniel C. Zilio, and Calisto Zuzarte. Recommending XMLTable Views for XQuery Workloads. Proceedings of the International XML Database Symposium (XSym), 2009.

University of Waterloo
Contact information: