Speaking at the Petroleum Industry Data Exchange (PIDX) fall conference in Houston this month, Pervasive Software CTO Jim Falgout spoke of the ‘big data challenge’ to oil and gas. Falgout outlined different approaches to big data. Analytic databases from Neteeza and Teradata are performant but ‘incredibly expensive to scale.’ ‘NoSQL’ data stores such as Cassandra, HBase and the open source Hadoop/MapReduce system offer ‘very cost effective scalability.’
Hadoop’s MapReduce programming model is powerful but hard to learn. Enter Pervasive’s ‘Dataflow’ programming paradigm which is scalable, supports iteration and is easy to learn. Dataflow offers a graphical programming environment for big data that can run locally or on a cluster. Dataflow programming is explained in a Dr. Dobbs article where Falgout traces its ancestry back to a seminal 1974 paper on parallel programming from Inria researcher Gilles Khan. More from Pervasive.
© Oil IT Journal - all rights reserved.