Building on the Edge: Evolution in Energy & Industrial Sectors

The Leading Edge: More Assets, More Data, Shorter Analysis-To-Action Loops

In the next 5 years, the number of connected and controlled industrial assets will outnumber the number of people on the planet by a factor of 3x¹.

From automobiles, drones, and manufacturing robots to solar inverters and wind turbines, the number of endpoints is quickly increasing — and the data-intensity of these assets is growing exponentially². While this trend is not new, recent advances in computing are augmenting how that data can be turned to valuable insights and actions.

Low latency requirements and bandwidth constraints are giving rise to shorter feedback loops and bringing about a new class of control applications for power generation, smart grids, autonomous vehicles, and industrial processes.

We see three fundamental shifts as industrial customers evolve to realize immediate value from operational data:

1. Accelerated improvements in computing power, power usage effectiveness³ and data center energy consumption, both in hyper-scale and micro-scale closer to the source

2. Novel communication technologies that provide low latency, high bandwidth & secure networks for real-time & computationally intensive operation of distributed assets

3. Increase in computing and analytics performed at the “edge”; that is, on the asset / sensor itself

Together, these developments can be particularly important for energy & industrial companies digitizing their operations.

Energy & Industrials: The Edge’s Perfect Match

A few seconds can mean a world of difference when managing power equipment or controlling autonomous vehicles.

We believe this transition to “edge computing” is necessary to enable use cases that demand real-time action, especially in industries requiring constant up-time.

To this end, edge computing can have a significant impact for:

energy, manufacturing, telecommunication,& transportation companies.

These sectors require low latency, high security & control, and often manage remote assets with limited bandwidth. The windiest areas in the United States are in remote areas of the Midwest, where wind farms grapple with poor network access (forget 5G, even 3G access can be a challenge).

Edge computing can limit the need for data “round trips” to and from the cloud.

Processing critical operational analyses at the edge effectively makes big data smaller, lowering a network’s storage and bandwidth requirements. A more distributed & networked system residing close to or on operational assets also limits single point of failure risks. Some assets may even analyze and operationalize data, while limiting cloud communications to critical information. As autonomous vehicle technology evolves, analytical processing and decision-making will likely reside in the AV itself — with remote telemetry limited to failure-mode data and periodic backhaul of historical operations.

Importantly, corporations are investing to digitize and secure a large and growing number of distributed assets. Nearly 80 million or more than half of all utility customers in the U.S. have smart meters, and nearly 20 million homes have smart thermostats⁴.

The digitization of “edge” assets is well underway, but now the question remains—

how will companies collect, analyze, and operationalize that distributed intelligence?

Our Focus: Computation, Communication, and Control

The transition from central to decentralized (echoing the journey from mainframe to personal computers, to client-server, to cloud) is clearly cyclical, and we believe it will be gradual. On this journey, we are most excited about companies building digital solutions in the Computation layer, the Communications layer, and the Control layer.

Computation Layer:

• Energy efficiency: improvements in data centers and beyond, including co-locating data centers with abundant renewable energy supply
• Workload arbitration: orchestrating hardware at the data center and allocating computing tasks between the edge and the cloud

Communication Layer:

• Machine to X: whether dealing with machine to machine, machine to cloud, machine to edge, or otherwise, asset-agnostic and scalable communication protocols… “APIs for Assets”
• Security: ensuring the security of data transfers and integrity of controls; monitoring and enforcing compliance across network

Control Layer:

• Enabling platforms: platforms to enable development and deployment of “edge-native” applications
• Edge applications: asset management and system optimization applications relying on advanced analytics to better orchestrate assets at the “edge”

We believe organizations building “horizontal” use cases serving multiple industries focused on enabling customers to “do more” (rather than simply lower costs) will win.

‍

REFERENCES

[1] Source: Statista, using estimates as of 2016

[2] Source: IDC predicts a 10x increase in annual data generation by 2025, to 163 zettabytes per year

[3] Power usage effectiveness (PUE) is a ratio that describes how efficiently a computer data center uses energy

[4] Source: EIA

‍