Skip to main content

Intel's Ronald Kasabian attempts to break big data down

What the heck is "Big Data," anyway? Along with the "Internet of Things" and the "Cloud," the term is one of those buzzwords that's probably spoken to impress folks more often than it's fully understood.

Intel is here to try to explain it all.

The first thing to understand is that the amount of data being generated by a wide variety of connected devices and machines is growing exponentially, to the point where Intel is expecting a 10x growth in the amount of data we produce today by 2016, according to Ronald Kasabian of Intel's Datacenter and Connected Systems Group.

The key point here is that 90 per cent of that data growth is expected to be what's called "unstructured data." That's information coming in from a wide variety of sources using different protocols and translations, which doesn't naturally mean a whole lot in the context of other sources of information, Kasabian said.

"This is data streaming in from video cameras, from manufacturing systems, from satellites, from social media, and more," he said. The challenge is to "correlate this information, using the cloud and using Intel's data assets to get insight into what it means," at which point one can begin to provide useful analytics, implement machine-learning techniques, and ultimately create value, he added.

"The big value of Big Data is the ability to correlate all these types of information coming in to create insight into what it means. The insight is the important thing, not the size of the data itself," Kasabian said.

Intel's view is that there are three key areas where Big Data analysis can provide value for businesses and consumers, which Kasabian outlined. It can help drive operational efficiency as with traffic optimisation and Smart Energy programs, create targeted, location-based marketing opportunities aimed at mobile device users, and advance security and risk management via things like insurance claim fraud detection, he said.

The challenges to making Big Data analysis work include the difficulty of correlating all those data streams, but also the enormous task of making sense of even single streams of extremely complex sources of data, Kasabian said.

Nevertheless, Intel is pushing forward to get even more complex in its treatment of data. For example, the chip giant is working with an HVAC systems manufacturer to process data from multiple systems rather than on a system-by-system basis for "a more aggregated approach to the Internet of Things," he said.

Within Intel itself, various projects are underway to better utilise Big Data and drive value. Those include the use of data analytics to better manage Intel's reseller channel, improve malware detection, and reduce product testing times. The last project expected to save $20 million (£13 million) for Intel in 2013, according to Kasabian.

"What do people think of when they think of Big Data? Well, it's big and it's interesting. But what people sometimes forget to talk about is what we can do with it and how it can drive value," he said.