There is much interest across the telecoms industry in edge computing. The key rationale behind this is that telcos – through their increasingly virtualised distributed network assets – are in a unique position to push workloads closer to devices, reducing latency and/or data volumes to centralised clouds, and thereby enabling new experiences and use cases, while enhancing existing ones. It is also becoming apparent that 5G is going to need all the help it can get, including from edge compute.
The following article is extracted from STL Partners’ new report on telco edge compute.
Enterprises are adopting edge compute. After years of centralising workloads in public and private clouds there is complementary demand emerging for more distributed compute. This is good news for telcos as it shows that the time is ripening for them to turn their ambition to edge cloud. Telcos – already building out telco clouds to support their virtualised network functions – can exploit their connectivity, unique network APIs and an existing distributed real-estate. Telcos should be a unique position to play a role in distributed and edge computing ecosystems.
Telcos’ edge cloud opportunity covers multi-access, enterprise, IoT “edges”
While there is much hype surrounding “the edge”, telcos’ perception of the edge opportunity is still evolving. This is in large part because there are several types of edge compute initiatives emerging from within operators:
- Multi-access edge compute (MEC): From a telco network-centric perspective, the edge resides on telco fixed and/or mobile network facilities, which distributed core datacentres, central offices, network aggregation points, eNodeB or base stations. There are a handful of standards initiatives underway, each related to specific application domains.
- Enterprise edge (also known as Intelligent Edge, Enterprise Edge, Virtual Network Services, Virtualised Managed Services): This is what many enterprises are referring to when they use the term “edge compute” or “edge cloud”. This refers to (potentially telco-managed) compute workloads that reside on an enterprise’s premises to keep data and processing local, e.g. for compliance or latency reasons, and potentially in combination with processing carried out in the telco network (e.g. hybrid uCPE-vCPE) or in the centralised clouds.
- IoT edge: Analytics and local IoT control functions can be carried out in an on-premise gateway to reduce the amount of data sent to the centralised cloud or – again – for latency or compliance reasons. Backhaul savings and autonomy in the event of network unavailability are key here.
- Device edge: The end-device (such as IoT devices, surveillance cameras, autonomous cars etc.) is also sometimes regarded as a part of the edge computing definition. This definition can extend to consumer devices such as Wi-Fi routers, smartphones. (we need to give some examples here like Alexa, Echo, smart homes etc… )
Telcos should individually seek to maximise co-ordination of their own edge compute initiatives across different domains (e.g. mobile edge, IoT, enterprise, device etc.), identifying and aligning technical and commercial approaches into shared platforms, ideally under a single business initiative. While “multi-access edge” is what most telcos think of when it comes to edge computing and edge cloud, telcos need to think across the different edge concepts (many already offer enterprise edge services), as they will be substitutive or complementary to telco edge cloud/compute.
Telcos should not restrict their edge ambition to the network. They should also consider compute resources that they manage on customer premises to be part of their edge infrastructure and edge service opportunity. For example, this means that telcos with a significant business customer base should seek to play in Enterprise edge.
Most importantly, telcos need to offer a joined-up vision to their customers and co-opetitors.
Telcos should plan to position existing (and potential future) edge capabilities within a wider cloud market context. The reality is that the big global cloud players (AWS and Azure in particular) have the established presence and have well-funded plans to expand on this. They also represent de facto standards for developers and application providers. Telcos can nonetheless carve out a role and co-exist successfully with the big cloud players. Telcos should seek to achieve this by working with – not competing against – the hyperscale players.
Edge compute remains a “potential” rather than proven opportunity: success will depend on applications running on distributed “multi-cloud” environments in a simple, seamless way. This means competing by co-existing with and complementing public and private clouds.