A new generation of databases that use SQL and a distributed file system like Hadoop are hitting the market. These NewSQL databases connect data professionals and business intelligence applications using SQL to a scalable data storage backend that isn’t slowed by data caching.
Monte Zweben, the CEO and co-founder of Splice Machine, said the market is ready for the right SQL database that can scale for a new era of applications that require tens of thousands of transactions per second and be ACID compliant. These databases are interactive, allowing for real-time iterative queries that new big data applications require. Products like Splice Machine, NuoDB and VoltDB, are trying to tackle those problems.
Zweben said technologies like MapReduce were created to solve specific problems, like massive search indexing. MapReduce and Hadoop developers left SQL out because it was crucial to the problems they were solving at the time. Now computer scientists are learning how to adapt those technologies to solve other problems, and adding in SQL is a natural evolution.
In this interview with Data Informed Staff Writer Ian B. Murphy, Zweben discusses why SQL has held on as a crucial technology, how Splice Machine uses HBase and HDFS to connect distributed data to a SQL engine, and why NoSQL databases are supposed to mean “not only SQL” but rarely take advantage of SQL.
Related articles on Data Informed: