This is the first article in a two-part series examining the edge computing opportunity for telecoms operators and hyperscale cloud providers.
Fixed and mobile operators find themselves at an inflection point, both in terms of their network assets and their balance sheet. In order to meet evolving user demands, they are investing heavily in next generation infrastructure including, with an aspiration to achieve monetisation in the future.
Differentiation through performance has been identified as a potential driver of this monetisation. It means catering to applications and use cases which exhibit a vast array of requirements, from low latency to high reliability and large capacity, and structuring business models around the different tiers of performance.
However, in order to deliver the anticipated leaps in performance enabled by next generation networks, telecoms operators need to fundamentally alter how traffic moves across their infrastructure. With 5G, for example, the latency improvements brought about by a new Radio Access Network (RAN) and air interface alone may be insufficient.
Defining the edge for telecoms operators
The edge is a difficult place to define because it connotes different things to different people. Even when examining the edge purely from the perspective of a telecoms operator, a breadth of interpretations exist.
Nonetheless, the consensus is that the edge spans any distance between end-devices including smartphones and IoT sensors, and the hyperscale cloud. For telecoms operators, the edge is only a subset of this hierarchy with a model such as Multi-Access Edge Computing (MEC).
Under a “telco edge” architecture, they may be able to position edge compute infrastructure, which encompasses computing, storage and networking functions, within or outside the public network. Think about the opportunity to leverage and locate at existing assets such as macrocell sites, street cabinets and network aggregation points in the access and core network.
To further minimise the physical distance over which traffic must transit, there is also an opportunity for the telco edge to be present on-premise at an enterprise. The decision regarding where to locate will be based upon a multitude of factors, and will likely be very much influenced by competition from players such as hyperscale cloud providers.
In addition to competition, the virtualisation roadmap of each telecoms operator will be an important consideration. This will dovetail with current software-defined networking (SDN) and network function virtualisation (NFV) efforts that are taking place under the umbrella of 5G investments, enabling network functions to be virtualised and distributed.
When implemented across each layer of infrastructure, from the core, transport and access networks to access points and premises, telecoms operators can exploit the benefits of the edge (low latency and reduced backhaul costs) while maintaining the advantages of the traditional hyperscale cloud (massive scalability and mobility).
The edge revenue opportunities for telecoms operators.
Despite being enveloped in hype, there is still uncertainty as to how telecoms operators can make an edge computing model commercially viable. This is acutely relevant to consider amid a climate of relatively stagnant revenue growth and increased capital expenditure on fibre and 5G infrastructure.
As alluded to earlier, the looming competition of hyperscale cloud providers (among others) is also key to note because it means that telecoms operators will only be a part of a much wider edge ecosystem. The share of the opportunity that they can obtain will depend on their edge business model and the ability of this model to cater to use cases.
Edge business models for telecoms operators such as that of a systems integrator, or co-location and infrastructure-as-a-service (IaaS), should be built around the benefits of caching and processing data close to the source.
The backhaul and interconnect cost savings achieved by transforming how transport networks are used, for example, may enable telecoms operators to reduce prices for edge customers and spur the development of radically new pricing models that take required transport distance into account (to offer lower pricing for processes that can be localised at the network edge).
Each business model will require varying degrees of upfront investment, risk tolerance and timescales in order to be successful. Some models will position the telecoms operator as an enabler and low down the value chain, while others will position it as an end-to-end solution provider and high up the value chain.
Whether the telecoms operator is most suitable to becoming an enabler or an end-to-end solution provider will depend on a number of variables. Enabler-type business models may work best where third-parties are already providing applications and, therefore, need to coordinate with many telecoms operators to exploit the benefits of edge IaaS.
The use cases enabled by these business models include virtual content delivery networks (CDNs), an edge cloud wholesale service, connected vehicles and AR/VR applications. Notably, however, millisecond latency may not always be demanded with every use case, and operators should be careful not to place too much emphasis on trying to attain this.
By contrast, business models that revolve around the telecoms operator being an end-to-end solution provider may be most desirable where mission criticality is of paramount concern. They will likely be required to meet service-level agreements (SLAs) for performance and comply with stringent privacy policies.
Examples of mission-critical use cases that can be enabled by an end-to-end solution provider business model include edge-enhanced content in the form of VR for live sporting events, AR for competitive gaming and analysis of CCTV video surveillance at the edge.
Conclusion: The edge complements the cloud
Edge computing is still in its infancy, both in terms of the underlying ecosystem and commercialisation efforts. For telecoms operators, there is an opportunity to secure a share of the market by developing a suitable business model that caters to the requirements of use cases and presents a revenue opportunity.
However, the telco edge should not be positioned as a competitor to existing hyperscale cloud solutions but, rather, a complement. In fact, it may be most prudent for telecoms operators to partner with existing players such as Amazon (AWS), Microsoft (Azure) and Google (Cloud) to develop solutions that can be adopted via a cloud platform.
Verizon, for example, is following this very strategy with Amazon Web Services’ Wavelength and its mmWave 5G network. Other players, like American Tower and Rakuten (which is creating an end-to-end fully virtualised cloud-native network in Japan), are also dipping their toes in the edge space.