Sensing a greater need in big-data analysis tools, IBM will invest US$100 million to research advanced large-scale analytics, the company announced on Friday.
IBM also said it will have 20 new service offerings to help customers quickly analyse petabytes of data.
“We think it’s a good time from a research perspective to double down on how we think about big data and how people can get actionable insights from it,” said Rod Smith, IBM vice president of emerging technologies.
The money will be used to investigate ways of advancing software, systems and services to better analyse data. “Analytics are terrific, but if you can combine analytics with specific problems and expertise around certain problems, then you can define outcomes in unique ways,” Smith said.
While the IT industry already has plenty of business intelligence and analytics software, tools are still lacking for analysing large amounts of unstructured data, Smith said. About 80% of the data in an enterprise is unstructured, meaning it is not easily categorised in databases.
Smith said many IBM customers are collecting large amounts of data through sensors, either software- or hardware-based, but do not have ways of analysing the results. “From a research perspective, we can ask, ‘What does that mean, and what can we do in these new areas to help them?’ ” Smith said.
IBM will use existing locations for the research rather than open new centres. “Projects will be evaluated across the globe,” Smith said.
The service offerings being developed will include a range of analysis capabilities, said Steve Sams, IBM Global Technologies Services’ vice present of site and facilities services.
One new tool, called Cloud Workload Analysis, will allow organisations to summarise their IT workload characteristics and identify which jobs could be moved to the cloud. Another tool will allow users to provision their own storage, rather than go through a storage administrator, reducing the time storage administrators spend managing resources.
IBM also announced a new software package that it said can help analyse petabytes of streaming data and derive results in under a millisecond, through the use of Hadoop.
The software, a combination of the InfoSphere BigInsights data analysis framework and InfoSphere Streams analysis tool, can quickly analyse text, video, audio, click streams, images and other forms of data. IBM claimed its software package can analyse data up to 350% faster than other offerings on the market.
In a recent survey of 3,000 chief information officers, IBM found that 83% agreed that analytics and business intelligence would play a pivotal role in their strategic growth.
