The Case for Big Data Analytics – And the Infrastructure to Support It
Big data analytics -- and its seemingly enormous infrastructure requirements -- hits an IBM System z sweet spot.
[Editor's note: This article was originally published in the SHARE User Group's President's Corner blog. Reprinted by permission of SHARE.]
By Renee Ferguson
If you were at SHARE's annual user conference in San Francisco in February, the "big data, big needs" conference track comes as no surprise. In fact, the topic probably comes as little surprise even if you weren't able to attend the event, because the discussion around big data analytics, and its seemingly enormous infrastructure requirements, can be heard (or read) everywhere these days. Fortunately, it hits at a System z sweet spot, according to Doug Balog, general manager for System z at IBM.
Balog delivered this year's keynote address at SHARE San Francisco. In it, he talked about key IT trends -- big data, cloud computing, security, and mobile -- that are occurring in the larger IT landscape. He spoke about how those trends map well with where System z traditionally has been, and where it's going in the future: enabling fast, secure, embedded analytics into a platform that is secure and, much like the cloud, virtualized and hybrid.
When I asked Balog where IBM's customer base typically is at with these trending topics, particularly big data analytics projects, he was frank:
"At the beginning, to be clear," he said. "That's an exciting place to be."
The reason? IBM has brought to market new technology that enables users to significantly minimize the time spent on real-time embedded analytics.
"We're seeing client improvements in operational analytics of 100-times to 1000-times improvement," said Balog. "When you get that kind of technology innovation, that is a fundamental breakthrough to allow clients to do things they never thought of before."
Balog points to two examples where users have been able to cut analytics processing time (enabled by System z, DB2 on z/OS and Netezza):
- Fraud detection: IBM's DB2 Analytics Accelerator enables real-time analytics for real-time fraud detection (think someone that isn't you standing at an ATM stealing your data -- and your cash)
- Online sales: While a sale is being transacted, System z-embedded analytics looks for the next-best action to increase sales value by offering something the buyer might be interested in -- while they're shopping.
Mark Anzani, vice president, portfolio and technical strategy at IBM, talked about the challenges and opportunities of big data analytics at SHARE San Francisco. The challenge: there is a hellish amount of data to contend with (two billion devices connected to the Web last year; three million e-mail messages a second; 200 to 300 million tweets a day). The opportunity: "The perspective on your customers is a wonderful thing."
The volume of data coming into the enterprise can shift dramatically based on even small changes, according to Anzani. He gave the example of utilities that have implemented smart meters, which sample electronic consumption every 15 minutes. Although the data provides much greater insight into patterns of use, it's also leaving some major utilities stuck, the proverbial deer in the headlights.
"I just had conversations with two utilities," said Anzani. "Guess what? They're not doing anything with the extra information of smart meters, other than aggregating it all and still feeding it into billing systems -- because of the complexity of having to do the analytics."
Barriers to adoption of big data analytics, according to Anzani, include:
- The time and expense to prepare data impacts ROI
- A lack of skills and experience applying analytics to data and business processes
- Traditional application vendors offering tight integration to data and business processes
- Pre-packaged and "good enough" solutions providing competitive entry points with line-of-business users
Citing a 2011 Forrester study that the former System z CTO said is still relevant today, Anzani mentioned that businesses are focusing application development and integration in three main areas: analytics, mobility, and social methods of collaboration.
"If as a company, based on the nature of your infrastructure and provider of technology, you're not considering the ways in which this is going to change the nature of your systems design and the management of how things get used, you will very soon see that data pilfering away to other platforms where those operations and those workloads can be run," said Anzani. "That's why it's very important to stay focused and understand the ramifications all the way through technology and the utilization methods of it."
Organizations are seemingly taking note of IBM's big data analytics infrastructure message. IBM saw its biggest year ever last year for the mainframe, adding 180 new customers -- only one-third of which were in growth markets.
Fourth quarter results mirrored that success, with the mainframe getting much of the kudos for IBM's earnings wins. From a recent MarketWatch article:
One interesting dichotomy in IBM is that, even though the biggest portion of its revenue now comes from its global services business, it still sells and upgrades the mainframe, which was once its main bread and butter.
New product introductions in its System z series were "very successful," IBM's Chief Financial Officer Mark Loughridge told analysts. Overall, its hardware business was down 1%, but excluding retail store solutions, which IBM sold, hardware was up 4%.
I asked Balog what he attributed the stellar growth in System z sales to. His response: "Clients are valuing the reliability and maturity of System z, but also the new capabilities around analytics and the cloud.
"When you can talk to a client about helping them attack total cost of ownership for a whole bunch of Linux x86 platforms and bringing those workloads under a much more efficient system that can run thousands of virtual machines that can save them roughly 50 percent of operating costs, that's pretty compelling."