Oracle looks to clear air on big data

IT vendor touts the need to establish the correlation between unstructured and structured data, rather than focus only on how to capture unstructured data like social media.
Written by Eileen Yu, Senior Contributing Editor

SAN FRANCISCO--A company's big data strategy should revolve around the correlation between unstructured and structured data, and not simply identifying how to capture unstructured data.

Speaking to Asian media at the OpenWorld conference here Wednesday, Christopher G. Chelliah, Oracle's Asia-Pacific vice president of Exadata and strategic solutions, said only 5 percent of data today is structured. The other 95 percent lies in unstuctured data primarily from the emergence of social media networks, customer information from telcos and utilities such as call history, messaging logs and usage trends, and information services such as traffic data, weather information and stock indices.

Such data is being generated very quickly and "goes to waste", Chelliah noted. Most organizations are not collecting the data because of the sheer volume, the velocity or speed at which it is being generated, and variety.

If harnessed properly, however, big data can drive new meaning for businesses, he said. Citing a McKinsey study, he noted that if captured and processed, big data can increase the value of the U.S. healthcare industry by US$300 billion per year and increase the retail industry's net margin by over 60 percent. It can also decrease assembly costs in manufacturing by 50 percent.

Not surprisingly, much focus today has been placed on how to capture and organize unstructured data.

However, this is only 20 percent of the problem, Chelliah said, adding that a lot of big data vendors were too focused on this.

To really captalize on big data, organizations need to establish a link between unstructured and structured data and have the ability to visualize this in their business environment, he noted.

"There's no point going through all of my Facebook feeds, LinkedIn connections or tweets if you cannot connect that to your list of frequent flyers---which is structured data--for instance, if you were an airline," he explained. "It's that last-mile, the correlation between the 5 percent of structured data and 95 percent of unstructured data which is the key part in getting value out of big data. We believe that's the next wave of growth."

Big data adoption in Asia-Pacific
Asked if the interest in big data had translated into actual deployment in Asia-Pacific, Chelliah said he is seeing customers implementing big data technology, with early adopters in industries such as telecommunication and retail because of the sheer volume of data they need to deal with. He also pointed to manufacturing as another early adopter.

One of the barriers to adoption was the lack of proper understanding of big data. "You don't know what you don't know," he said, adding that Oracle had been addressing this through conversations with customers and engagement with its partners.

Roger Li, senior vice president of technology business unit at Oracle Asia-Pacific, also noted the many different definitions of big data today, similar to how cloud was loosely defined when it first emerged as a market trend.

He said it was more important to address the challenge businesses face today. "And the challenge is there's a lot of talk and ideas, but it's how you turn these ideas into a structured way [that's important]," Li said.

Eileen Yu of ZDNet Asia reported from OpenWorld 2012 in San Francisco, United States, on the invitation of Oracle.

More coverage from Oracle OpenWorld 2012 on ZDNet:

Editorial standards