The discussion about big data to date has been primarily focused on the volume, variety or velocity of data. We are making the assumption that readers are familiar with big data and, indeed, with the concept of the Internet of Things (IoT) so we will not be discussing the “three Vs”. However, now that big data deployments are beyond the trial stage and into deployment it is time to consider some of the issues that arise when big data implementations transition beyond skunk works and into general-purpose use and, in particular, the issues that arise when organisations are integrating their mainframe system of record alongside big data implementations. Note that by “mainframe” we mean an IBM z/OS environment, for which we are using “mainframe” as shorthand.
More or less every major organisation in the world has a mainframe at the heart of its enterprise and it is critical that big data deployments are viewed from that perspective, rather than treated as isolated efforts that are distinct from the mainframe environment. As the system of record, big data must ultimately (if not sooner) link back to the mainframe. What we will discuss in this paper are some of the issues surrounding these big data deployments and how they might be resolved within the context of a mainframe environment.