There's a new school of thought starting to emerge in the world of analytics that stresses the value of analyzing streams of data in real time.
The biggest proponent of this approach to analytics is IBM, which is using a platform it developed to run streaming analytics as a cornerstone of its Smarter Planet initiatives.
After initially making InfoSphere Streams available in beta late last year, IBM today announced the general availability of version 1.2 of the platform. This version of the platform adds a higher speed messaging transport, more predictive analytic applications and a built-in failover capability, said Nagui Halim, a researcher and product director for InfoSphere Streams at IBM.
The basic idea behind the InfoSphere Streams is to analyze streams of data in real time in order to predict an outcome, and then use that information to alter that outcome if it proves to be undesirable. The example IBM gives of the technology in action is how InfoSphere Streams is being used in Stockholm to alter traffic patterns. But the InfoSphere Streams could just as easily be used to track blood flow in a patient in order to establish what changes in patterns are early indicators of the onset of a particular disease. Other applications include developing analytics around various types of trading systems and applications that could track e-commerce patterns on the Web in real time.
InfoSphere Streams has its own run-time environment and requires mastering a new dedicated language IBM developed to build applications on top of the platform. To facilitate that process, IBM is making InfoSphere Streams an element of the analytics training that it is funding at schools such as Fordham University and the University of Ottawa.
While the concept of analyzing streaming data as part of business intelligence applications has been around for a while, IBM is trying to create one of the first dedicated frameworks around this discipline. It is one thing to use predictive analytics to gather information after the fact to come up with some data that suggests what might happen next based on past events. Predicting what will happen next based on live streaming data is a whole other level of analytics that has far greater potential to change the way people actually work and live every day.