Calculations for Rent


Powerful, affordable data analytic computations are emerging from the cloud

By Katie Kuehner-Hebert

Through a growing array of renting options, community banks can now access increasingly sophisticated analytical computing power, according to technology analysts. More platforms that offer vast computing capacity from “virtual machines” to perform millions of complex calculations—that would typically crash in-house software programs like Excel—can be affordably tapped for data analytics and other number-crunching projects, they say.

“One of the great benefits of cloud computing is that even the smallest institutions can embark on big projects without incurring a significant upfront capital cost,” says Daniel Latimore, senior vice president for banking at Celent, a Boston-based research and advisory firm. “If they run certain [data analytic computing] projects, they could use the ‘pay-per-drink’ model and get results that five years ago would not have been possible.”

As software-as-a-service solutions provide software to companies on the cloud, “infrastructure-as-a-service” and “platform-as-a-service” solutions provide the necessary virtual backbone for companies to develop their own software, so they don’t have to invest millions for in-house servers. A community bank’s IT staff can buy time on such services for one-time projects—such as comparing the interchange fees of several prospective ATM vendors based on actual transactions—or continuously rent time on cloud computing services for ongoing calculations, such as credit-risk simulations for a steady stream of loan applications.

Some community bank core providers are developing their own data analytic cloud services, and “it’s only a matter of time before it becomes more prevalent,” says James O’Neill, a senior banking analyst for Celent. He adds that DH Corp. “is probably most along the trail of generating community bank adoption,” with its Compushare C3 solution, an ICBA Preferred Service Provider cloud services program.

O’Neill stresses that these emerging “true” cloud computing services are different than the data warehouse analytics packages that are offered by bank vendors, in that the cloud computing services allow for a real-time capability in monitoring a bank’s transaction flow. Moreover, current data analytics packages are priced on a traditional per-seat-license basis, while emerging cloud services will allow for pay-per-use economics, in which a bank pays for the specific level of service it uses.

O’Neill outlined one scenario of possible cloud computing power available to community banks in the report, “Cloud-based Financial Services: A Banker’s Guide,” published in November: Suppose a community bank whose ATM processing agreement was up for renewal. The bank is faced with the challenge of comparing the debit interchange income arrangement of its current processor with those proposed by three other processors, using the bank’s actual 1.7 million annual ATM transactions as a foundation for comparison. Each transaction carried a unique merchant category code that was tied to a specific interchange rate, and each ATM processor maintains on average 20 different standard categories of possible merchant interchange rates.

In this hypothetical scenario, the community bank would try to compare the interchange tables of the four processors using Excel spreadsheets, but the program would likely crash given the massive number of calculations that would need to run each time an interchange rate changed or a new one was added. However, if the bank transferred these calculations to a cloud computing platform that uses numerous “virtual machines” and the latest cloud data analytics technologies, the work could have been completed in seconds.

Several public cloud services are now offered by a number of providers, including Amazon Web Services, Microsoft, IBM, Google and Salesforce. According to a December report by Forrester Research Inc., “The Forrester Wave: Enterprise Public Cloud Platforms, Q4 2014,” by analysts John R. Rymer and James Staten, public cloud platforms can cater to the varying types of software developers within most IT departments.

“Thus, for many enterprises, the best choice of a public cloud platform will be a platform—or a portfolio of platforms—that addresses multiple types of developers working with several languages and frameworks,” Forrester’s report says.

Pricing for such cloud services would be based on which types of software a bank would want to use on the cloud, and for how long, O’Neill says. Community banks could also choose which service to use based on how much their IT staffs were willing to code themselves, versus leveraging ready-made tools to help them develop their applications.

“If you have a large data set, particularly one that is being constantly updated, the data analytics services you can get from the cloud create a much better environment for getting the job done than trying to do it on a machine yourself,” O’Neill says.

Katie Kuehner-Hebert is a financial writer in California.