About Associate Professor Uwe Roehm
Uwe Roehm aims to improve performance for data management in systems software by innovative design and careful evaluation.
Associate Professor Uwe Roehm is a world-recognized researcher on database management systems and on data processing in sensor networks. He developed freshness-aware scheduling (FAS), a novel approach to replication management in massive clusters that allows clients to trade data freshness for query performance. The innovation of this approach is the introduction of an explicit freshness limit for accessed data as a quality-of-service parameter for clients, and the adoption of a layered transaction model for asynchronous replication management. The FAS protocol allows for different versions of data, and uses the freshness limit to provide faster service for clients with lower freshness needs, while guaranteeing 1-copy serialisability. He developed a suite of query routing protocols—cache-approximation query routing—that allows the routing of queries to cluster nodes, depending on the approximated cluster state, thereby increasing throughput significantly. These routing protocols are completely non-intrusive and hence highly generic and system-independent. Their routing decisions are based solely on the observed workload, using dynamic approximations of the data accessed, derived from SQL statements. With these protocols, it is possible to coordinate large database clusters with more than 100 nodes.
Associate Professor Roehm completed a PhD at ETH Zurich in 2002. He then worked as an IT consultant in Germany before joining the University of Sydney in 2004. He has been a reviewer for leading database conferences and journals such as International Conference on Very Large Databases, the ACM Conference on Information and Knowledge Management, the ACM SIGMOD Conference on Management of Data, the Conference on Innovative Data Systems Research, IEEE Transactions on Database Systems, and Information Systems Journal. He has been a visiting researcher with Microsoft’s SQL Server group. He is Chief Investigator on an ARC-funded Discovery Project and also on a Linkage Project funded by ARC and Microsoft.
More information about Uwe Roehm