Big Data Not Yet Big

It's the new kid on the tech block. If you are not lipsyncing it at your networking meets, you might be considered not clued-in. Vendors are going overboard with it and the analysts are putting big fancy numbers on its market progression and call it the next disruptive thing in enterprise computing.

It's Big Data!

Data deluge is not a new thing but over the last few years this deluge assumed mission critical proportions. Those who mined it, hit the gold mine in terms of understanding their consumers and their businesses better. Big data is all about your strategy to store and use data from all sources and having the ability to get useful business insights out of it.

We will not go into more of what big data is as the CIO community is well in the know about it. The real question is: Data deluge has existed for years, so why this whole excitement about it now? The answer probably lies in that right now we are at the intersection of data explosion and BI/analytics; the former happening because of new and divergent sources of data and the latter because of its ability to come up with useful business insights.

The new and divergent sources of data include social networks, blogs, and many other hitherto untapped sources like real-time data from machines, unstructured forms of data, et al. For this we need a new framework to manage the process and that is broadly called big data. It's not just about analytics, it's about a whole ecosystem needed for data modeling and outcomes - higher computing power, better BI tools, state of the art enterprise class storage, and of course lots of IT spend. No wonder vendors of all hues are going all out.

Ground Realities

Let's get up close and personal and try and figure out the CIO mindset right now on big data. We spoke to a handful of large enterprise CIOs and asked them about their perspectives on big data and how they see that manifesting and impacting their IT set ups.

Venkat Iyer, CIO at Wockhardt sets the tone by making an honest assessment of things by saying: "big data is when you have data of 100 TB and above, and which cannot be easily managed with the use of existing analytical tools. Big data also means that the composition of data where more than 50% of data is semi or unstructured. Seriously most of the companies need not lose their sleep on this as other than insurance, banks and mobile companies not many have Big data. Few media companies would like to analyze the unstructured video content which runs into petabytes of data but there are no tools available."

Iyers's views rekindle the adage 'Old wine in a new bottle'. And it looks big data is exactly that, and with aggressive evangelizing the vendors are trying to re-innovate and re-invigorate storage and data management and BI. As the term big data aptly implies, the data volumes have to be humongous for enterprises to seriously consider big data framework and go for that granular level of insight and discovery.

While it's hype right now but there is consensus that data is growing and very soon all enterprises, irrespective of verticals, have to mine that data more intelligently. The emphasis senior CIOs put forward is on a proper due diligence of the need requirement and, the ability to see through the hype and industry jargons is a critical factor in understanding any technology at its hype phase.

Quips Arun Gupta, CIO at Shoppers Stop, "In my assessment big data right now is partly hype and partly opportunity. CIOs need to be aware of what the new buzzwords mean to their business so as to have a rationale conversation with the business teams. There will always be a next big thing and with the widespread awareness of such trends through non-IT media, everyone is talking about it. CIOs have the opportunity to take the lead and educate their business leaders on the need, alignment and impact if at all."

Gupta also feels that the vendors should play responsibly and partner with progressive CIOs to create demystified scenarios that can be used for better discussions and engagement. "I believe that storage technologies are well placed within the Indian enterprise; the new solutions may not apply equally to everyone and that is the qualification that the CIO needs to do in collaboration with the vendor to separate the hype from reality," avers Gupta.

Leading CIOs say that few enterprises in India are already experimenting with big data and most have yet to reach consensus on ownership and how to leverage the opportunity. Take the case of Shoppers Stop for instance in which Gupta is trying to put a big data strategy in place.

Says Gupta, "We have started engaging some solution providers to assess impact to business. Over the coming months we plan to pilot a few use cases to determine the strategy and direction. The technology is also yet nascent; so real impact and benefits will take a while to get off the ground."

Creating a big data strategy and implementing is not that easy as well. Putting this in perspective Venkatesh Natarajan, special director, IT at Ashok Leyland says, "The solutions to handle big data are expensive. Hence it will be prudent to use these solutions to store only data that has importance to the business."

"The following can be considered: One can resort to de-deduplication so that multiple copies of the same data do not reside on central storages. Moreover most of the unstructured data are generated by individuals these could be photographs, movie files, presentations or other such files. Since multiple versions of such files, could be there on the local disks of users, and could also find a place in central storage areas, it would be good to formulate and implement policies which necessitate removal of older versions to save space and certainly self-discipline of users will help a lot," he adds.

The Early Adopters: Telcos and Retailers

From an India specific point of view and globally as well, the early adopters of big data are verticals like telecom, retail and to an extent BFSI and government are some of the early adopters. Like telcos, the high-end research sector in bio and life sciences are also using big data for last many years. In fact it's nothing new for them as they have been into big data for quite some time. That does not mean it's not relevant for other verticals, but they just don't need such computational analytics and their data size and plurality are not that big.

Quips Wockhardt's Iyer says, "Big global retail chains of the world are yet to come to India. Indian retail business is still in its infancy. At the global level some research organizations are thinking about recording the heart beat across millions of people around the world and look at patterns. This will create big data. We are a long way away."

Speaking to Amrita Gangotra, director, IT, Bharti airtel reveals how verticals like telecom are leveraging big data for the last many years. She says, "As a telco we have been into big data for many years. The scale of our operations makes us a very data driven organization. We use a combination of technologies to slice and dice the big data and we are able to create successful business outcomes and customer lifecycle management. Moreover we are able to map customer preferences and be able to devise innovative plans and drive revenue performance as well. For us, big data is extremely critical."

However Amrita believes that the whole phenomenon of big data has been hyped up and it has just got a new name. She attributes that to the growing data management challenges and the enterprises in non-telecom verticals and a growing realization to mine the data more effectively for business outcomes and just for statutory requirements.

Yet another company that makes good use of big data is Reliance Communications. According to Alpana Doshi, CIO, RCOM, "The CIOs approach on big data depends on the acceptance of the real fact. How the company thinks and what are their visions and strategy in handling these complex problems. As far as RCOM is concerned we are already in the foray in managing the spurt. Reliance has been actively involved in big data solutions due to sheer size of our databases. We have adopted Multi Parallel Processing DB to store CDR and unstructured data and perform analytics on it. We have implemented storage virtualization and unified storage in a single box."

Yet another vertical that makes good use of big data is retail. Yet again Indian retail space needs to achieve that kind of scale one sees in the West. But just to get a sense of what really is big data at retail, sample this: Wal-Mart handles close to one mn customer transactions per hour and the size of its database is around 2.5 petabytes and growing. Available information suggests that the size of Walmarts database is equivalent to 167 times of all the material available in books in Library of Congress.

Looking at the Walmarts case, one gets a sense of how niche big data is right now and the kinds of enterprise which really need big data tools to slice and dice their data repositories. Says Anand Rajaraman, SVP, Global eCommerce at Walmart, "There is more data now in the world than a few years ago. And in retail this is largely about products and customers. The thumb rule to follow in a big data strategy is to focus on the problem that you want to solve rather on data. And find as much data as possible to address the problem, the more sources and data you have the better will be your positioning and strategy around it."

Beyond Hype

As we cut through the hype, the frenzy on big data becomes palpable. Data has become multifaceted and the boundaries of traditional data predominantly generated by employees and stakeholders are getting highly impacted by social media platforms. Moreover with models like BYOD getting rampant employees within and enterprises are connecting to the corporate network and creating more new forms of data across device form factors - tablets, smart phones, apps, multiple OS' et al - has created not only data deluge but plurality. And when the elasticity of the conventional analytical tools and BI solutions becomes static, big data solutions are the way forward.

So clearly most of the CIOs in India are taking the wait-and-watch approach when it comes to big data. And many believe that in the next 12 to 18 months it will become a major trend and morph into a key element of their enterprise computing framework. Nevertheless all of them agreed that they will be looking at a big data strategy at some point in time in the near future given that across enterprise data volumes are set to grow.

Up Ahead

Due to increased usage of data intensive technologies (Digital cameras, RFID, mobile devices, etc) the average size of all types of files have increased. This is expected to lead to an explosion of data in the unstructured area, which will be a challenge to handle, in the next couple of years. Reflecting on this, Venkatesh Natarajan says, "In the structured area (ERP and other applications using databases) also, businesses want to run complex analytics almost in real time. This is leading companies to look at technologies like in-memory computing. Thus the ability 'to handle big data' may become important to businesses in India, though not immediately."

There is no doubt that big data is hyped up, but at the same time one should not overlook the value and its potential. Too much information overload even in non-IT media on big data, is creating a surrealistic perception and masks many important issues relating to current data management practices, the need and the ideal candidates of big data, and how some of those learnings can be emulated by enterprises that really do not need big data but can leverage some of the data mining and analytical practices at a lesser scale for quicker output.

At the end of the day what this debate on big data is ushering in is putting the spotlight on BI, analytics and storage management practices in the limelight and providing bigger and better ways of mining that data. As Shashi Mohan, CTO and CIO, Polaris Financial Technology sums it up: "Most of enterprises deal with huge amounts of data and they are engaging in preliminary conversations on how to better utilize that data. Big data right now is evolving and my take is that in another year or so it will capture the imagination of CIOs in a big way."