Distributed Analytics Forecast: Partly Cloudy with a Chance of Fog

Distributed energy showcases the role of advanced analytics

Distributed energy continues to be a major focus for many utilities around the globe. That was certainly the case last week in San Diego at the electric utility industry’s major tradeshow for transmission and distribution, DistribuTECH 2017. As it was last year, one of the major themes of the event was how to build out and support an increasingly decentralized grid. Many solutions to support microgrids, energy storage, edge devices, and Internet of Things connectivity were demonstrated. In conjunction, discussions on edge communications and security were abundant. The thread that ties all of those components together is distributed analytics.

For utilities, what still seems to up for debate is how these distributed analytics can be most effectively deployed. Embedded edge analytics appear to be the current focus of many solution providers and equipment manufacturers. Several current use cases place an emphasis on limiting the amount of data used and processing it close to or on the source.

However, as electricity generation, delivery and consumption becomes more transactive, utilities will need to tap data from multiple points across the enterprise, and it won’t be just operational in nature. Customer data (social sentiment for example) and other enterprise information will be crucial. Edge analytics will continue to play a vital role, but they won’t be enough. Additional distributed analytics will be required. And the electric grid is but one example of how distributed analytics were proliferate across many industries.

For electric distribution, driving value from this broad spectrum of data will require utilities (or third-parties) to employ multiple layers of advanced analytics, including machine learning. Data will need to be tapped in north/south and east/west layers across the business and even into business and residential premises. That eventual reality suggests that fog analytics is the end game for utilities. With that in mind, let’s look at the different ways edge is incorporated into distributed analytics, including fog.

Cloud only: Edge data to cloud analytics

Edge analytics is common in some industries, retail and finance for example. However, until recently, analytics at the edge wasn’t possible in industrial settings due to a mix of cost, complexity, security, and technology barriers. Instead, edge-to-cloud analytics was the default starting point.

With this approach, edge data is sent to a cloud for storage and processing. Models can be created and data run through an analytics engine. Results ae shared via back-office system integration or through reports to help inform people in their decision making. In some instances, results can be sent back to the edge devices or, more commonly, to edge control systems to modify device performance.

Additional data from the enterprise or third parties may also be stored, processed and analyzed in the same cloud environment. If a use case supports it, that additional data might can be used for edge analysis. This centralized model provides the most power for processing, storage and analytics.

Local edge: Analytics computed locally

A second way to incorporate edge data is by using a local cloud or server. This method leverages the power of a traditional server architecture or centralized cloud. The edge gateway handles communication to the edge device, ingestion of data, and transportation of that information to the local server or cloud. The local server or cloud manages data readiness, storage, analytics, and visualization. Security, privacy, data-related cost, and regulatory constraints are often the reasons cited for keeping the analytics local.

Edge only: Embedded and edge network analytics

This instance excludes the cloud or any other centralized processing, even if it is local, and so it is the technical definition of edge analytics. The analytics are embedded within an edge machine or in devices delivered via an add-on edge technology. They can be supported in a nearby gateway. Data storage and analytics are done on the device or within a network of edge devices. This is beneficial when bandwidth requirements are an issue and the analytics are less complex than what is typically processed at a platform level.

Fog: Analytics distributed throughout the network

The most recent innovation related to edge data envisions intelligence distributed at any data node. This approach is referred to as fog analytics/computing.

Fog analytics operates on the principle that processing should be done where it provides value. In this instance, flexibility is the overriding functional requirement, as the analysis might occur at any layer of the operational footprint, extending to any horizontal or vertical point. Thus, a use case may use any mix of device-only, local, network layer, or enterprise analytics.

Learn more at this year’s ARC Forum

If you are attending this week’s 2017 ARC Forum and are interested in learning more about topics discussed in this blog, please join me for my session Best Practices for Implementing Analytics for Digital Transformation today at 4 p.m.

Leverage ARC to Build Your Business Case

If you are unable to join us in Orlando this year, ARC offers a full range of solutions designed  to help you understand and implement advanced analytics, including:

• Buyers planning guide: A great starting point to guide business case development.


• On-site workshops: A comprehensive session for you or your executive team that explains analytics and highlights key issues of the buying process.


• Technology selection support: We use ARC-designed functional requirements to help you plan, build, and execute customized request-for-information (RFI) proposals.


 

To learn more, contact me or an ARC client manager.