IBM on Wednesday will unveil the commercial version of its stream computing software, which is used for real-time analysis of data.
The commercial version of this IBM Research software will be dubbed System S. IBM will also establish the IBM European Stream Computing Center in Dublin, Ireland. In a nutshell, System S can analyze data on the fly because it doesn't store it. Previously, software needed to store data befor analysis.
Here's how stream computing would work in an enterprise:
The commercialization of streaming computing is part of Big Blue's analytics push. IBM argues that its streaming computing breakthrough can analyze various unstructured data sources in real-time for quick decision making. Although the obvious fit for streaming computing is the financial industry, IBM argues that the offering will apply to multiple industries such as law enforcement, retail, transportation and health care.
For instance, Uppsala University and the Swedish Institute of Space Physics are using the streaming computing tools to track space weather's effects on various signals. TD Securities is using System S for trading algorithms. And Big Blue and the University of Ontario Institute of Technology are using it to track subtle condition changes in premature babies.
Steve Mills, SVP of IBM Software (left), and Dr. John Kelly, SVP of IBM Research show off stream computing.
IBM argues the aim of System S is to deliver "perpetual analytics" and algorithms that can find the data in a haystack that lead to a competitive edge in real time. IBM argues that current analytics approaches are too slow. System S is designed to prioritize data, overcome data format woes and tap unstructured information from the Web, electronic sensors and video.
IBM said it will make System S trial code available for free so customers can try it out. The code includes developer tools, adapters and software to test applications. According to an IBM spokesperson pricing will range from $100,000 for a two server development installation up to several million dollars on large clusters with hundreds of nodes.
For now, System S can only be bought directly through IBM in English. The product is expected to be fully available in multiple languages and through IBM business partners in 2010.
According to IBM the streaming computing approach is one way to tackle a massive amount of data. IBM said that by 2010, the amount of digital information is expected to reach 988 exabytes, roughly the equivalent of a stack of books from the Sun to Pluto and back.