Ahead of the 2013 CIO Summit on 31 July in Singapore, K.R. Sanjiv, Senior Vice President, Analytics & Information Management Services at Wipro Technologies, shares his views on what need to be done before organisations can embark on the right route towards Big Data.
Photo: K.R. Sanjiv
Q: Can organisations build an infrastructure to take advantage of big data? How can data be organised into coherent parts so that purposeful distillation can be conducted to derive useful information to guide organisations ahead?
K.R. Sanjiv: For handling big data, the organisation requires an architecture which is significantly different—in fact, drastically different, from what they have conventionally. They have to address issues like shelf life of big data, quality of the data, and volume. The data may not be stored like in a database. And data attributes are fundamentally different. Fitting a big data solution into an existing architecture would fail. From a single-tier architecture for data warehousing, it should be extended to a three-tier architecture for big data distillation. Out of 90 percent of data out there, only 10 percent is relevant for extracting.
What is the business processes to evaluate, what to do with the huge pool of data, what to extract and what could be relevant are all issues that must be addressed. A layered approach to big data has to be built first. On one side is to handle raw data, and on the other, all the logic (rules) to extract the information. Half of big data is actually outside the organisation. To that extent, this layer has to claw and get data from the outside, and the challenge is to siphon off the noise to make sense to business.
There are two dimensions where Wipro can help, because big data is more than just handling data; it is not the end goal. In conjunction to analytics and visualisation, we have to understand what data to consume. Then only there's business sense. Looking at whole stack, and marrying it to analytics, then the business can see the benefits, like increasing revenues, or reducing costs.
Also, new disruptive technology can provide a lot of opportunities for innovation. Customer segmentation, fraud detection, product categorisation, retail… these are some of the direct benefits of business analytics. Not just that, new revenues for monetising and leveraging big data are also possible.
Where are the pain-points before reaching the next big thing on data processing and analysis, in particular, predictive analytics?
There are a lot of pitfalls when you talk about big data and how organisations struggle with it; one is technology itself. Put in another way, just getting the tools alone isn't going to help—it's then a solution looking for a problem. How should organisations go about doing it? First is to look at, say, their top five business problems, identify where their priorities should be, relook at their IT investments, and then work backwards on how big data can be leveraged.
Sign up for Computerworld eNewsletters.