A leading provider of self-service kiosks, automated teller machines and financial security services, required a cost effective framework for analyzing machine generated data from over 200,000 globally distributed devices. Part of a larger cloud initiative, the firm’s CIO turned to big data technologies to help accelerate the company’s goal of “setting the pace of innovation and product development.”
The client’s CIO initially tasked an internal team to launch a data analytics service that would use predictive analytics to determine when ATM components would likely fail. This information would help maintenance and service teams improve ATM uptime, a key satisfaction measure for customers. The internal team quickly realized their initial predictive maintenance solution (based on Oracle’s relational database) would not scale technically or financially. Requiring a cost effective and scalable solution, the client turned to Cloud Technology Partners for its expertise in predictive maintenance, big data frameworks and PaaS.
Cloud Technology Partners provided the architecture and support for a scalable, global cloud platform that will enable the company to perform large-scale basic and predictive analytics and derive valuable insights into its operations. Collaborating with the client’s team, Cloud Technology Partners first gathered key technical requirements and provided best practices, a data analysis model, use cases and a reference architecture. In the second phase of the engagement, Cloud Technology Partners evaluated big data vendors against key requirements and assisted the client in selecting the matching technical components including a core big data platform, an advanced analytics tool, and development tools. Cloud Technology Partners used its standard capital planning model to help the client make trade-off decisions around the cost / benefits of different SLA’s, hosting models (internal vs. cloud), cloud providers, and server profiles. The result provided a full year’s budget request broken out by line item. As part of the final evaluation, Cloud Technology Partners built a functional proof of concept using HortonWorks HBase with standard analytics and machine learning.
Upon approval of the capital request, the big data framework will be incorporated into the client’s PaaS, enabling their global delivery teams to deliver innovative new analytical services and is expected to reduce ATM maintenance costs by $5-10M per year. As different data sets and analysis approaches are fielded, additional cost savings are expected.