Report forecasts power requirements for edge use cases

According to a forecast model built by Tolaga Research, the mobile consumer segment is expected to make up the largest edge infrastructure footprint (as measured in power consumed) at nearly 17,000 MW, followed by the residential consumer segment at 10,843 MW. Significant footprints are also expected for the communication network operator, enterprise IT, retail, health care, automotive and smart cities segments, each between 3,000 and 5,800 MW, according to the report.

Defining the Edge-and how much power will be needed

The report emphasizes that the tech community is collaborating on coming up with a common translation of edge. The State of the Edge group, which is described as a member-supported educational organization, uses the LF Edge glossary to define terms used in the report. While the device edge incorporates end-point devices and gateways, and enterprise on-premise servers, the infrastructure edge consists of edge equipment at access network sites such as radio base station tower sites, pre-aggregation sites such as street-side cabinets, and aggregation and central office sites.

Forecasting power footprint, by use case

Following years of extensive tech trials, smart cities will see a booming growth, as many of their services make for great edge use cases in large-scale roll outs such as smart-lighting, traffic management, smart buildings and public venues and services for public safety and utilities, according to the report’s findings.

The report’s forecast calls for the global IT power footprint for smart city-related deployments of infrastructure edge equipment is estimated at 158MW in 2019 and growing to 3052MW by 2028.” While in 2019 more than half of the power footprint is used on smart building and public safety solutions like video surveillance, in 2028 they will account for 36.9 percent and respectively 24.8 percent of the global smart city footprint, while the rest will be assigned to traffic signaling, smart lighting, public venues and utilities.

Source: State of the Edge 2020

The report’s authors predict that traditional workload deployment with on-premise equipment will continue to become less common. Enterprises can leverage the infrastructure edge to migrate device edge workloads such as data management, privacy and security to off-premise locations and public cloud without losing the benefits of on-premise equipment, they suggest. Workload deployment at the infrastructure edge might be a better option than cloud computing architecture for latency sensitive and bandwidth intensive services, but edge and cloud are not mutually exclusive, the authors note.

AWS, Azure offer Early signs of edge growth

AWS and Microsoft are two of the main public cloud providers in the market with an extensive cloud product offering that can be set up either in the device or in the infrastructure edge. These companies promise to deliver added-value through AWS’ Outpost and Microsoft’s Azure IoT Edge products which easily integrated with AWS and Azure public cloud. In line with customer demands, the two products can now be utilized on both device edge and Infrastructure edge implementations. The report predicted the two would seek out affiliation with communications network operators (CNOs) and regional data center providers to make them feasible for Infrastructure edge implementations.

Earlier this year, Microsoft and AT&T announced they would combine 5G Network Edge Compute (NEC) technology with Microsoft Azure cloud services into AT&T network edge locations to extend core cloud applications to edge computing environments in the enterprise. Shortly before the report’s publication, AWS announced at its re:Invent 2019 conference in Las Vegas that it was partnering with Verizon to create Wavelength, a solution for ultra-low latency applications for 5G. Wavelength merges AWS compute with storage services at the edge of telecommunications providers’ 5G networks to help companies build apps for mobile end-users and devices with single-digit millisecond latencies.

Luana Pascu contributed to this article.

Article Topics

data center  |  edge computing  |  edge infrastructure  |  power consumption  |  State of the Edge