The big picture: What's going on with edge?

This is an extract from TM Forum's new report How to build and operate at the edge, download it now for the full insight.

Edge computing is an evolution of public cloud that is happening regardless of how, where, when or whether communications service providers (CSPs) find a place in the value chain. Public cloud providers like Amazon Web Services (AWS), Google Cloud Platform and Microsoft Azure are driving the change by extending computing services to the edge of the telecoms network to manage sustained growth in traffic and improve customer experience. They are also making speculative investments in edge computing capabilities to meet the anticipated demand for IoT services.

The telecoms industry has an important role to play in edge computing because CSPs provide the connectivity that edge computing requires. The powerful combination of 5G connectivity and edge computing can unlock new latency sensitive consumer and B2B services, such as gaming and telemedicine. Operators also own passive infrastructure such as telephone exchanges that could be suitable for housing edge computing equipment in some cases.

CSPs would like to bring together edge and network strategies. However, they are still in the early stages of deploying 5G, and the market for computing services is not mature enough for them to do so. This makes it difficult for them to determine where to collocate edge network functions.

Our research found that many operators are unsure whether their companies will pursue an edge computing strategy at all, and respondents in every group cite lack of solid use cases as the most significant barrier to moving forward.

Common themes

Briefings with executives reveal several important findings that are consistent across companies and regions. For example:

Getting started is difficult – operators have put a lot of thought into their strategies around edge computing but are struggling to commit to and start executing a clear plan of action. Most activity comprises trials or proofs of concept, usually involving partnerships with public cloud providers and sometimes including third-party content providers and applications developers. Only a handful of Tier 1 CSPs have launched their own edge cloud services in select locations.

Public cloud providers are in the driver’s seat – they are setting the timetable for CSPs’ investments in and commitments to edge computing. Operators are turning to cloud providers as partners for their digital and network transformations, and cloud providers are urging them to deploy edge computing capabilities as part of these partnerships. CSPs are being forced to decide whether to proceed with their own edge computing plans and determine how they could affect partnerships. They are also trying to figure out the right balance between forging preferred relationships with cloud providers and adopting a multi-cloud strategy.

Deciding where to deploy edge computing capabilities is key – CSPs are conducting detailed assessments of where to deploy edge computing sites (either alone or with a partner) by mapping out the locations of their own B2B customers, factoring in where key verticals need connectivity.

5G timing is everything – CSPs want to combine deployment of edge computing and 5G core network capabilities. However, most operators are still in the very early stages of deploying the 5G core. Furthermore, many of the technologies and capabilities that could justify deploying a highly disaggregated core network, such as open radio access networks and network slicing, are still being developed and may not be ready for widespread deployment for two to three years.

CSPs are developing local and national strategies – in the B2B market, many requirements are to serve local campuses or factories, so operators likely will target these customers with a cloud provider partner. Nationwide capabilities are more likely to be part of a CSP's 5G core network strategy.

Private LTE and 5G networks will impact edge computing – in most cases organizations operating private mobile networks (for example, a port or manufacturing company) will take a hybrid on-premises and public cloud approach to managing IT workloads. This could remove CSPs from the revenue chain if industries build and operate their own networks and connect directly to the public cloud.

Latency is a major driver for the deployment of edge computing, though not the only one – the consensus is that 20-30 millisecond “in-air” latency will suffice for most applications. For small countries this means that edge computing may be required in only a handful of locations. Other drivers include delivering a better experience, capturing more of the enterprise market and reducing the cost of transport.

Using the local exchange is not a significant revenue opportunity – none of the CSPs interviewed for this report believe that renting space in their local or regional exchanges to hyperscale cloud providers will be lucrative. Furthermore, it is by no means certain that these facilities can be used to collocate edge computing gear because they lack space and power. These conclusions from interviews are broadly consistent with the results of the online survey. However, it is worth noting that while our briefings were largely with Tier 1 CSPs that have already revealed their edge computing ambitions, the majority of operators deploying 5G are smaller players in their respective markets and do not have a strong enterprise business to drive edge computing.