“Edge computing” can mean many things, ranging from computing and storage on a device itself, through enterprise on-premises computing, a network or telecommunication (telco) operator’s network edge, and even computing in the core of a network operator. All of these could be considered the “edge” compared with the global-scale computing capabilities offered by some data centers in a particular region.
In the telco world of 5G, “edge computing” generally relates to the pushing of computing processing capabilities to the edge of the network, close to users. Multi-access Edge Computing (formerly Mobile Edge Computing, or MEC) is the name of a set of specifications used for this technology in a cellular environment, for example in a mobile base station, or if not there, then at least closer to the user than typical Internet data centers. Technical standards for MEC are being developed by the European Telecommunications Standards Institute.
MEC provides a distributed computing environment for application and service hosting. Applications can also be exposed to real-time radio access network (RAN) information. The key element is the MEC application server, which as a multi-tenant system provides computing and storage resources, connectivity, and access to RAN information. Individual applications run as virtual machines (VMs) on the application server.
The technical benefits of MEC are proposed to include better user quality of experience for low-latency, high-bandwidth applications being hosted or processed at the edge, lower load on the backhaul network due to caching at the edge, or potentially the ability to dynamically move workloads to and from the edge in “containerized” platforms as required.
The first and primary user of edge computing resources is the telco operator themselves – either for virtualizing network functions or for their own operational and billing support systems. However, telcos also see an opportunity to host third-party workloads that can take advantage of the low latency, close-to-the-user benefits of edge computing.
Commercially, telco operators believe that such third-party hosting would enable them to re-insert themselves back into the end-to-end nature of the Internet value chain, and to derive value from new use cases and new innovative market segments that need high-bandwidth, low-latency connectivity – for example connected cars, industrial IoT (Internet of Things), video delivery or video analytics, AR & VR, or cloud gaming. There is an opportunity for telco operators to rent MEC resource to Internet companies, and enable operators to move up the value chain by providing a “scarce resource that they could monetize. Telcos are therefore interested in the commercial opportunity for MEC, and want to understand the technical requirements for the new use cases it enables, in particular in 5G networks, where edge computing is a core part of the 5G specifications. The commercial possibilities of edge computing are an important part of the business case for operators investing in 5G.
An early-stage investigation of telco operators to date suggest that there have been two options, potentially talking past each other:
- Option 1: Telco companies see edge computing as a revenue opportunity to host and serve cloud products, services and workloads from their own infrastructure close to users, generating revenue for the telco while doing this.
- Option 2: Many cloud providers already encouraged the use of standard Cloud constructs for edge to compute resources – for example containers – to enable dynamic movement of telco (or third party) applications and workloads between edge, core, and web-scale computing resources – generating revenue for themselves when those applications end up on their own platforms.
Regarding some potential benefits and risks of these options, the following comparison can be made:
A reality check would show that so far, although many use cases of MEC have been proposed, there have been very few if any real-world applications that have adopted the MEC architecture. In fact, there is orders of magnitude more computing power available at the true “edge” – in smartphones, PCs, and even IoT devices, than in any edge computing capabilities proposed by telcos. For example, Dean Bubley made some great foresight in one of his articles- “MEC and Edge Computing are overhyped and underpowered”. Bubley claimed that “in-network edge-computing architectures, such as MEC, will become more important and that there will be almost no application that runs only in the network-edge as this will be used just for microservices, as a subset of a broader multi-tier application. The main compute heavy-lifting will be done on-device, or on-cloud so that collaboration between edge-compute providers and industry/web-scale cloud will be needed.”
Bubley further asserts that “given these developments within the MEC architecture, it looks like rather than becoming distributed quasi-Amazons- in the sense of running image-processing for all nearby cars or industry 4.0 robots in their networks, linked via 5G- few developers will want to write directly to telcos’ distributed-cloud APIs on a standalone basis, with or without network-slicing or 5G QoS mechanisms.”
Within the landscape of these computing resources, there may also be some unintended consequences such that a self-driving car’s computer could be used to offload tasks from the network, rather than vice versa. Who knows what the future will offer!