How IBM Champions tackle big data, part 2 of 2
Delaney Turner 270003RQ8K Delaney.Turner@ca.ibm.com | | Tags:  information-insights ibmsoftware
0 Comments | 4,937 Visits
Welcome to part 2 of my interview with IBM Champions Alex Philp, Founder and CTO of TerraEchos and Ivo-Paul Tummers, CEO of Jibes. In part 1, we explored strategies and algorithms these innovators are using to transform volume, variety and velocity into business value for their clients. Here, we discuss the broader enterprise-wide repercussions of big data and how they make the vital link to business analytics.
You can download the audio version here.
Let's talk about that multi-disciplinary approach to big data – at what point does a big data strategy change from an IT challenge to a business imperative?
Alex Philp: I really like what Ivo-Paul said about the value of being able to look at all of these sources from multiple perspectives. We’re all realizing that the game today is analytics. It’s delivering the right analytics at the right time, and applying them at the right time in your workflow. What’s driving our customers in government, in telecom, in oil and gas is the idea that they need to optimize their business processes and their workflows. They’re looking for the right analytics against the right sources to produce the right answers.
There’s actually an amazing corollary between government mission and what a CEO, CFO or CTO needs to see on a daily basis with their operations. We’re not talking so much about specific hardware or software. We’re approaching customers with a solution-based, interdisciplinary approach. We say, “Let’s talk about how we’re going to save you time and money so that you can stay competitive in getting the right information to your customers.” Time savings is a great way to show ROI.
Ivo-Paul, where are your clients on this realization?
Ivo-Paul Tummers: First of all I completely agree that saving time is the best ROI. I also think the timing of IBM’s big data strategy is splendid because I think we’re on the verge of exactly that moment. Where IT departments are already struggling with the challenges of data integration, growing data sets and integration with other systems in the value chain ‘the Business’ is not fully aware yet of the potential.
As a result we see two type of customers. The first are those where IT is still fighting an uphill battle with small and often disconnected initiatives and projects resolving operational issues. The second are those where the board has adopted the concept and see the potential and need to change. This results in more ambitious projects with clear objectives on returns and tactical and strategic advantages.
Let’s look at the connection between Big Data and Business Analytics more closely. Have your clients altered their IT infrastructures or processes to make these capabilities work better together?
Ivo-Paul: Analytics is the most important business goal for a big data strategy. More data provides better analytics outcomes. Beside that, big data should eventually be brought back to a small and meaningful answer. Our Big Data team operates under the label of “Data Analytics & Applied Intelligence.”
Alex Philp: We can’t find enough mathematicians. I can’t find enough people to support the algorithms throughout the stack. The other challenge is that a lot of people are looking to do analytics and big data in the cloud, and synchronizing analytics across a high-volume, high-velocity networks is a challenging situation. That’s not trivial.
We try to look at what we can help customers achieve with a Bayesian approach to statistics, or perhaps a Markov approach when we’re working with incomplete data. We’re trying take complex approaches to data science and mathematics and turn them into usable commodities to talk about “Analytics as a Service.”
Our customers don’t want all the analytics all the time; they want it on-demand so they can look horizontally through the data, pull out the key pieces, do the correlation and then derive some type of function to get a better look at where they’re going. It’s that steering that we’re trying to package up for our customers in a very cost-conscious world. If we can’t show them savings, we need to show them efficiencies.
That’s an incredibly tall order. What IBM technologies are you using to put together?
Alex Philp: I agree with Ivo-Paul that IBM is the world leader in this space. That’s clearly why our company is so involved with IBM. Customers have a lot of options within the IBM big data platform. We focus primarily on the Infosphere streams component – on the real-time, high velocity side. Then we focus on scaling that with the right combination of IBM hardware. We take streams and IBM hardware, we optimize both, then layer in our analytical packages and tools to get data in and out. This could be Hadoop, Netezza, DB2, even a customer’s call center system.
The final piece we’re working is metadata. We have so much video and audio to work with now, but it’s not always relevant or practical to be processing and reprocessing it all the time. We’re trying to create and store the metadata – it might have a temporal component or a spatial component - and develop metadata analytics that complement the IBM components that we’re leveraging every day. We think that’s best of class.
Ivo-Paul, what are your thoughts on the tools and technologies you’re using?
Ivo-Paul Tummers: Alex mentioned metadata; you also need to look at the master data aspect, which is also covered by IBM. Other new technologies are the Hadoop platform, and in memory and/or columnar databases. People need to be trained to handle this, not only from a development perspective, but also from a system administration perspective. For these teams it sometimes looks like a revolution instead of an evolution.
It will take time to adapt to a Big Data platform. The main idea is to keep it simple and small, what should sound strange when talking about big data. But paramount is that the organization needs open-minded employees to work with this new technology. They should be the advocates for other departments.