To open and close this bar, use alt+/

Test of Time Award

September, 1, 2019 - NYU Abu Dhabi Assistant Professor of Computer Science Azza Abouzeid has won the prestigious 2019 Test of Time Award from Very Large Data Bases (VLDB).

VLDB Test of Time Award paper is selected from the VLDB Conference from 10 to 12 years earlier that best meets the “test of time”. The committee evaluates the impact of the paper in practice, e.g., in products and services. Impact on the academic community demonstrated through significant follow-through research.

VLDB is considered the premier publication for research in data management and data systems. The Test of Time Award is the highest honor given out by the VLDB endowment.

In 2009, Abouzied’s research integrated two then competing technologies, large scale data processing systems such as Google's MapReduce and Apache Hadoop, and parallel database management systems, to build a hybrid data analysis system: HadoopDB. HadoopDB delivered the best features from both technologies: it provided the high SQL query performance and efficiency of parallel database systems and the scalability, fault-tolerance and flexibility of large scale data processing systems. Therefore, it could easily scale to clusters of thousands of machines despite their high failure rates and varying machine performance. 

Computer scientist Azza Abouzeid accepts the Test of Time Award from Very Large Data Bases (VLDB).

Abouzied’s research was one of the early transformative works that brought about deep architectural changes in the design of database and data processing systems. HadoopDB was the foundation of a startup company, Hadapt, that had significant industrial impact. Currently, many popular big data systems are hybrid large-scale data processing / parallel database management systems such as Impala, SparkSQL, Drill, Hawk, Hive, Presto and others. HadoopDB drove further follow-on research such as Abouzied’s work on invisible loading. Invisible loading incrementally moved data from distributed file systems into an optimized structured data storage and query execution layer without user intervention and with minimal overhead. HadoopDB received over 1,000 citations.