BigData SQL: Spark

Apache Spark is an in-memory distributed computing platform for fast general-purpose data processing and computation. Spark supports different types of computations— batch, interactive, stream processing, and iterative machine-learning algorithms—on the same framework. One of the distinguishing features of Spark is its ability to run computations in memory and store intermediate results in memory without going to disk.
Because varying workloads can be supported within the same framework and processing engine, Spark is a very convenient and productive framework to use from code development, code maintenance, and deployment perspectives. This is so because developers, devops, and dataops teams have to learn only one framework—its internals, nuances, and best practices—rather than different frameworks for different workloads.
This also has been the major reason for adoption of Spark as a distributed processing framework.