Article Post

Does IT architecture matter as much as energy analytics?

Data Entry

Analytics can deliver on the promise of new opportunities by revealing the insights needed to successfully pursue them. Customer-centric utilities need analytics for better targeting, higher response rates and improved campaign ROI. If you need to manage 2 million records with 400 different variables to test against that data, inadequate infrastructure (or IT architecture) will result in sluggishness at best and completely losing the struggle at worst.
Building on the past and planning for the future

Recently, my team and I discussed enterprise architecture with a U.S. utility's chief IT architect. We discussed how important it is for analytics to integrate into their existing infrastructure, add business value on top of it and even fit into their future planned infrastructure. It is not just about storing data efficiently, you must store it a format which allows analytics to be run in a timely manner.

Without a system that has been engineered and designed to scale, the breadth and depth of your analytics can't deliver the payback you expect. After all, if you can't deliver information to decision makers in a timely manner, it doesn't matter how advanced your analytics team is or how well your software can solve business problems.
To illustrate, below is a diagram that shows a subset of the data management, analytics and reporting options you should look for if you're developing an architecture for turning data into information and quickly delivering critical insights throughout your organization.

Preventing information lag

The utility industry is adopting analytics as a way to make more informed decisions. Will your infrastructure present barriers to analytics deployment? Maybe you need analytics on a PC, a server (SMP or MPP) or a group of servers (grid). You may need the ability to move important data or analytic processing workloads to where the data "lives" using in-database or Hadoop accelerators. Many organizations need an overall faster processing environment via an in-memory approach.

For best results, you may even need to move data and analytics processing out onto the edge using event stream processing for continuous analysis of events as they occur. This incremental updating of information enables the real-time analysis of trends to detect anomalies immediately.

If you have a smart architecture that scales to your analytical needs, any or all of the above may be used to capture additional value that could otherwise be lost through information lag.

This makes it easy to see why analytics is "the ultimate renewable resource." With the right infrastructure improvements, analytics can better reveal patterns and anomalies, identify variables and relationships and predict future events so you can select the best courses of action.


This article is specifically addressing needs with IT Infrastructure to support complex analytics, and does a good jov of describing what is needed. But IT architecture at the application layer is just sa important. Without a service-oriented architecture modeled on the IEC CIM or Multispeak models; analytics at the IOU and large-medium muni/coop level will always fall short of expectations.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.