Reality check: edge computing is not yet mature, and much is still to be decided
Edge computing is still a maturing domain. STL Partners has written extensively on the topic of edge computing over the last 4 years. Within that timeframe, we have seen significant change in terminology, attitudes and approaches from telecoms and adjacent industries to the topic area. Plans for building telco edge infrastructure have also evolved.
Within the past twelve months, we’ve seen high profile partnerships between hyperscale cloud providers (Amazon Web Services, Microsoft and Google) and telecoms operators that are likely to catalyse the industry and accelerate route to market. We’ve also seen early movers within the industry (such as SK Telecom) developing MEC platforms to enable access to their edge infrastructure.
In the course of this report, we will highlight which domains will drive early adoption for edge, and the potential roll out we could see over the next 5 years if operators move to capitalise on the opportunity. However, to start, it is important to evaluate the situation today.
Commercial deployments of edge computing are rare, and most operators are still in the exploration phase. For many, they have not and will not commit to the roll out of edge infrastructure until they have seen evidence from early movers that it is a genuine opportunity for the industry. For even more, the idea of additional capex investment on edge infrastructure, on top of their 5G rollout plans, is a difficult commitment to make.
Where is “the edge”?
There is no one clear definition of edge computing. Depending on the world you are coming from (Telco? Application developer? Data centre operator? Cloud provider? etc.), you are likely to define it differently. In practice, we know that even within these organisations there are differences between technical and commercial teams around the concept and terminology used to describe “the edge”.
For the purposes on this paper, we will be discussing edge computing primarily from the perspective of a telecoms operator. As such, we’ll be focusing on edge infrastructure that will be rolled out within their network infrastructure or that they will play a role in connecting. This may equate to adding additional servers into an existing technical space (such as a Central Office), or it may mean investing in new microdata centres. The servers may be bought, installed and managed by the telco themselves, or this could be done by a third party, but in all cases the real estate (e.g. the physical location as well as power and cooling) is owned either by the telecoms operator, or by the enterprise who is buying an edge-enabled solution.
Enter your details below to request an extract of the report
Operators have choice and a range of options for where and how they might develop edge computing sites. The graphic below starts to map some of the potential physical locations for an edge site. In this report, STL Partners forecasts edge infrastructure deployments between 2020 and 2024, by type of operator, use-case domains, edge locations and type of computing.
There is a spectrum of edge infrastructure in which telcos may invest
Source: STL Partners
This paper primarily draws on discussions with operators and others within the edge ecosystem conducted between February and March 2020. We interviewed a range of operators, and a range of job roles within them, to gain a snapshot of the existing attitudes and ambitions within the industry to shape our understanding of how telcos are likely to build out edge infrastructure.
Table of Contents
- Executive Summary
- Reality check: edge computing is not yet mature, and much is still to be decided
- Reality #1: Organisationally, operators are still divided
- Reality #2: The edge ecosystem is evolving fast
- Reality #3: Operators are trying to predict, respond to and figure out what the “new normal” will be post COVID-19
- Edge computing: key terms and definitions
- Where is “the edge”?
- What applications & use cases will run at edge sites?
- What is inside a telco edge site?
- How edge will play out: 5-year evolution
- Modelling exercise: converting hype into numbers
- Our findings: edge deployments won’t be very “edgy” in 2024
- Short-term adoption of vRAN is the driving factor
- New revenues from MEC remain a longer-term opportunity
- Short-term adoption is focused on efficient operations, but revenue opportunity has not been dismissed
- Addressing the edge opportunity: operators can be more than infrastructure providers
- Conclusions: practical recommendations for operators