Analytics on the edge using IBM Cloud Pak for Data

By | 3 minute read | May 19, 2020

As we grow smarter and more sophisticated, thanks to rapidly enhancing technological innovations, enterprise data management and analytics have to keep pace to ensure organizations continue to remain effective and data- and insights-driven.

Let’s not underestimate the sheer scale of the problem though. As existing systems of record are augmented by the explosion of data points generated by customers across a growing ecosystem of mobile, web, in-store and other points of contact, it is clear that the need to process data has grown beyond what the traditional data center can handle.  It is therefore no surprise that we are seeing an emerging trend growing in popularity: the extension of the enterprise cloud to the edge of the network topology.

As seen in the figure below, by “edge” we are referring here to the plane of computing that sits furthest away from the enterprise cloud or data center at a lower center of gravity: putting it closer to the source of the data.

Let’s not fool ourselves though; this problem is staring us in the face already and has resulted in many businesses being unable to utilize a large portion of their data. Some estimate this to be as high as 73 percent. And for those businesses who have started tackling this problem, a new challenge emerges – how do they analyze, learn and drive operational insights in real-time?

Traditional data management and analytics architectures call for this data to be moved to a centralized data center or enterprise cloud for analysis. This is not only expensive – think data movement and additional storage costs – but also increases risk and puts greater pressure on regulatory compliance and data security controls.

At IBM, we have recognized this modern-day data and analytics challenge and are building IBM Cloud Pak for Data as the enterprise data and AI platform to support. We’re supporting this new pattern of usage through Analytics for the Enterprise Edge. Cloud Pak for Data-driven Edge Analytics will allow businesses to analyze data on edge gateways using applications and analytical models created by developers and data scientists on the Enterprise Cloud/Data Center platform. It will help industries to make their businesses nimbler by providing analytical insights in near-real time while offering the flexibility to control what data gets sent over the network to the enterprise cloud or data center platform. The result: near-real time actionable insights while lowering costs and associated risks.

Overall, IBM Cloud Pak for Data driven Edge Analytics will help businesses in three areas:

  • Accelerate data monetization while reducing cost and latency – drive near-real time insights at the source; no need to move data to a central enterprise cloud or data center platform
  • Enhanced data security and compliance – Reduce unnecessary data movement, thereby reducing risks from unknown sensitive data
  • Extend the enterprise multi-cloud to the Edge – Cloud Pak for Data is designed to operate across multiple enterprise clouds, while seamlessly extending to the network edge

Cloud Pak for Data fits in synergistically with the IBM Edge Application Manager. The intelligent and flexible platform that provides autonomous management for edge computing. Cloud Pak for Data will provide the data and AI plane for the enterprise to build out advanced analytical assets and effect data management functions, while the IBM Edge Application Manager manages the deployment of these assets across endpoints, including edge gateways, where Cloud Pak for Data powered edge analytics would run.

IBM is also extending the value of Cloud Pak for Data driven Edge Analytics by partnering with strategic technology vendors such as Equinix and Eurotech, with a lot more to come as we continue to further strengthen the offering. We invite you to engage and partner with us on edge analytics, a pattern of enterprise computing that is here to stay.

Learn more about IBM edge computing solutions.

Accelerate your journey to AI.