Technologies never roll out as quickly as expected. Barriers to implementation arise, exuberance turns to pragmatism and numbers are restated. Not so with 5G. Omdia's "Mobile Subscription and Revenue Forecast," published in September 2020, reports that 5G subscriptions have grown by a factor of 15 over the past 12 months. They grew from 15 million in 2019 to over 230 million in 2020 and are on course to reach over 3 billion in 2025. That is 4x the rate of adoption of 4G, in terms of subscribers. As the industry began to feel the full impact of the COVID-19 pandemic and the global quarantine in 2Q20, 5G rollouts slowed — as did subscriptions — affected by stutters up and down the supply chain (including spectrum auctions).

As the end of 3Q20 approaches, however, carriers are regaining momentum and assessing the long-term, global impact of COVID-19 on network traffic, topology and usage patterns. These consequences have been well documented and include the rushed adoption of video conferencing tools by all generations, the change in residential traffic patterns and the overall increase in traffic, calls, over-the-top (OTT) video, live video, etc. The rollout of 5G — already fast — is now pulling edge computing with it.

In collaboration with Schneider Electric, F5 and Wind River, Heavy Reading asked over 90 global telco service providers about their plans for edge computing. In our report, "Edge Computing in Telco Networks: Gaining the Competitive Edge," Heavy Reading analyzes the choices these companies are making with their implementation — and how these implementations are evolving.

Technologies driving the telco industry evolution The survey results demonstrate that telco service providers are now deploying, or planning for, edge computing solutions. They are no longer at the nascent stage. Heavy Reading explored further by examining which technologies will affect telco service providers the most within the next two and five years. The responses reflect the way carriers tend to move — from the infrastructure, through supporting network technologies, and finally to the application enablement technologies.

Figure 1: n=91
(Source: Heavy Reading)

The response of "open and industry standard infrastructure" attests to the carriers' focus on containing cost, both capex and opex. However, the need for application and workload acceleration is encouraging bespoke white box-plus solutions with field-programmable gate arrays (FPGAs) onboard to siphon off compute-intensive processes, relieving the CPU cores.

Artificial intelligence (AI) and machine learning (ML) will also be critical to managing costs and supporting the flood of new 5G cell sites. When it comes to zero-touch provisioning and lights-out data centers, AI and ML will play a pivotal role.

The impact on existing business model
But what about edge computing itself? Heavy Reading asked survey respondents what impact edge computing was having on their business models; 40% responded that its impact was significant or radical. From what we have seen in previous research, that is an abrupt increase in the level of expected impact. In two years, the expected significant/radical impact grows to 70% of respondents. In five years, virtually all respondents (88%) believe edge deployments will have a radical/significant impact on their existing business model.

Figure 2: n=91
(Source: Heavy Reading)

When looking at the challenges facing the telco service providers in edge deployments, "cost to deploy" is the most significant barrier. It is the only one cited by more respondents as a "significant challenge" than as merely a "challenge." This barrier is followed by two other cost issues: "lack of a cost-effective operations and management model" and "lack of a business model."

Cost is not always the key concern of Heavy Reading's survey base. In fact, with a survey base weighted heavily toward network operations and technical staff, it is rarely the key concern. However, edge is all about cost. Telcos contemplating a robust edge infrastructure, highly distributed to the point where the edge sites can meet sub-1ms one-way latency demands, are looking at tens of thousands or even hundreds of thousands of locations. Cost, both capex and opex, is the key issue.

Figure 3: n=91
(Source: Heavy Reading)

"Lack of use cases" is fifth in a list of seven challenges (or fourth if going strictly by "significant challenge"). If Heavy Reading had conducted this survey at the beginning of the year, we would have expected this line item to rank much higher in the list of concerns. At that time, carriers were still shaking their heads at the parade of PowerPoints expounding on autonomous vehicles and remote surgery. Today, virtual radio access network (vRAN) and video processing/caching are important use cases. However, most telco service providers deploying edge computing today are being driven by general efficiencies in network operations and application performance and by user demand, particularly within vertical industries that require low latency response time for specific applications. These use cases are not as flashy as autonomous vehicles, but they justify edge deployment today. Both consumer and enterprise customers recognize the basic need for an always-on network.

Read the full report for more telco edge insights
Heavy Reading's findings are a good indication that telco carriers are poised for the business influx that edge computing will bring. To gain more in-depth details of telco service providers' perspective on the burgeoning edge market, check out our visualizer with key highlights and download and read the full report now.

This blog is sponsored by Schneider Electric.

— Jennifer P. Clark, Principal Analyst, Cloud Infrastructure and Edge Computing, Heavy Reading