So why is this important? Why is a new breed required? With Big Data, the architectures of the past don’t work well when having to deal with the size and growth of new and fast growing data (clickstreams, genomics, sensor data, etc.). This new breed focuses on performance. For many organizations faster results translate into better business results. Better performance can also mean less hardware. For example, we’ve had customers run faster with MapR on test clusters that are one-quarter the size of their production clusters. At the scale of Big Data with clusters of hundreds of nodes this results in big money.
Addressing the Big Data challenges requires a new approach. In the Hadoop space, this isn’t about wrapping an open source project with additional services or some management components. This requires fundamental changes -- low level architectural breakthroughs -- that drive improvements that are orders of magnitude better in speed, scale and efficiency.
MapR recognized several years ago that innovation was needed for Hadoop. It’s great to see this validated by a well-respected industry leader like O’Reilly.