Making the Case for an Intelligent Edge in an IoT World

Cloud computing gets plenty of attention in IT circles and among grid managers—it is hard to ignore when technology giants like Amazon, Microsoft, IBM, and others keep promoting their cloud solutions. But as the Internet of Things (IoT) concept gains momentum, new attention is being focused on intelligent edge and distributed computing.

This theme was prevalent at the recent Internet of Things World conference in Silicon Valley, where participants pointed out advantages that edge has over the cloud. Jesse DeMesa, a strategy partner at the venture capital firm Momenta Partners, said that a cloud-first or data center-first approach to IoT analytics will not work, and companies will ultimately move toward more autonomous systems. Many current IoT adopters, he said, focus on connecting, collecting, and storing data, while the “real value of data has a shelf life often measured in seconds.”

Getting the Edge on Costs

His point is well-taken. Looking back and analyzing large datasets for actionable insights via a cloud scenario does have value. But gleaning insights within seconds, or fractions of a second, at the edge and making immediate adjustments can be equally valuable, if not more so, when critical operations are at stake or the safety of nearby personnel is in play.

The need to rethink a cloud-first approach was emphasized for cost reasons by HarperDB CEO Stephen Goldberg. During a panel session, Goldberg said the bandwidth needed to push data to the cloud and the edge storage infrastructure is expensive, and ends up grabbing a significant share of the cost in an IoT deployment. He argued that a more distributed computing infrastructure, where edge devices already in place are doing as much computing as possible, is a more rational approach. This is true.

IoT vendors recognize this need for advanced intelligence at the edge. Recent examples of companies with new edge offerings include: SWIM EDX, which processes edge streaming data in real time; C3 IoT and Intel partnering on a new artificial intelligence (AI) appliance for optimizing applications that do not require cloud computing; and Edgeworx, a startup that builds software for edge gateways and micro data centers.

Edge Networks Get Sharper

The big tech players have noticed the growing edge computing need and the advantages of putting more and strategically useful computing horsepower there as well. Dell and Microsoft, for instance, have teamed up to create an integrated IoT edge platform. Amazon Web Services has updated its edge computing platform, called Greengrass, to incorporate machine learning capabilities. Similarly, Hewlett Packard Enterprise’s Aruba subsidiary launched a new network edge solution called NetInsight in March that uses AI to autonomously monitor corporate networks and optimize performance.

What this all means is that the edge of networks is getting smarter. It means companies deploying IoT solutions need a strategy that integrates edge, on-premise, or cloud computing architectures to take advantage of each for an enterprise’s or a grid operator’s own needs and applications. In some scenarios, a cloud architecture makes sense, or an on-premise solution. But more likely the complexity in these deployments will require a sophisticated blend of technologies. As my colleagues Richelle Elberg and Mackinnon Lawrence note in their Guidehouse Insights white paper From Smart Grid to Neural Grid, the future mature Energy Cloud will be based on technologies that integrate ubiquitous connectivity, cloud-based AI, and edge computing. That same type of integration of computing power will be needed for enterprises beyond the energy grid that seek to harness IoT as well.