With 5G now being rolled out commercially, communications service providers (CSPs) have a huge opportunity to capture new business thanks to unprecedented network speeds, capacity and massively reduced latency.
Given these performance benefits, it’s not surprising that operators are looking beyond enhanced mobile broadband and building their 5G business cases around a range of B2C and B2B applications and services that includes the likes of augmented and virtual reality, gaming, connected and autonomous vehicles, mission critical services and Industry 4.0, IoT and a whole host more.
Multi-access edge computing (MEC) is expected to play a pivotal role if anything like the full range of use cases projected for 5G is to be realised. In contrast to today’s centralized network architectures where the processing, storage and access to data are concentrated at the heart of the network, MEC distributes these functions out to the network edge, closer to the end user.
In a MEC network, localized content caching, real-time processing and management of data, and ultra-low latency connectivity are located at or driven from edge-network nodes or micro data-centres, all connected back to the central cloud, but with the flow of traffic between the core and the edge-network minimized in order to reduce latency, cut loading on the central network, and save on costs such as backhauling.
More importantly, by building this edge cloud out of a 5G network, MEC can bring into play a host of 5G network features to support new business opportunities in areas such as industrial automation, production and robotics. These 5G features include uplink classifier (ULCL) for dynamic traffic steering, network slicing for guaranteed end-to-end reliability, and session and service continuity (SSC) modes for seamless, ultra-low-latency connectivity. In addition, time-sensitive networking (TSN) techniques can be employed to guarantee low-latency.
Key to MEC’s success will be the creation of a converged ecosystem where connectivity and computing work together. Coupled with the extra speed, capacity and massively reduced latency of 5G, the edge cloud will enable differentiated service innovation through the convergence of network, cloud and devices. Upcoming developments in the standards for MEC will see work to address vertical industries as part of 3GPP Release 16, and the creation of an enhanced 5G architecture to support edge computing.
As the primary providers of network connectivity, CSPs are well-placed in the race to build the edge, and MEC’s innovative network architecture provides them with huge potential in terms of service creation and new revenue opportunities. A majority of CSPs have taken a business-driven approach to their transformation journey, preferring to adopt a full-stack model as the quickest way to provide them with the desired agility. The combination of network virtualization and telco cloud in this transformation journey is already pushing service providers closer to the reality of edge computing, and the move to a cloud-native, container-based reference architecture is building greater flexibility into the network and providing the capability to distribute resources closer to where they are needed.
In terms of the physical MEC infrastructure, operators’ existing assets such as edge or central offices can potentially work as a starting point for edge computing deployments, so overcoming the limitations on computing and storage capacity close to the network edge. Applications that require the caching or processing of large amounts of data at the edge may be better sited at a central offices or metro aggregation points. Longer term as 5G is rolled out, these sites can also be used to host virtualized core network components and associated services.
A fully implemented MEC scenario also comprises centralized orchestration and management of the MEC platform. Support for control and user-plane separation (CUPS) allows control-plane functions to be deployed centrally while user-plane functions can be flexibly deployed at the network edge. Application server functionality ensures that applications and services accessed via the edge nodes are seamlessly supported by the network. This will ensure service continuity and content synchronization for applications such as V2X and VR where users are moving at the network edge.
Enhanced hardware will be important for profitable new service creation and delivery, in order to help achieve the lowest possible cost-per-bit. Edge installations are likely to be space-limited but must support high traffic levels, hence compact hardware will be required. Flexible and scalable compute resources can provide an open, heterogeneous computing platform for service acceleration, capable of addressing applications, content, core network service processing and resource scheduling functions at the network edge.
Cross industry cooperation will be an important factor in order to create an on-demand, high-performance and open MEC platform. Industry consortia are already bringing together network equipment providers, chipset suppliers, providers of data centre and computing hardware, software developers and network test companies. Standards bodies and industry initiatives such as ETSI MEC and the Linux Foundation are also deepening their collaboration in the drive to define a common vision for edge computing.
A number of key questions about MEC remain to be answered, including which services and applications will require key MEC features such as ultra-low latency. Nevertheless, the incentive for edge computing to be deployed in 5G networks appears to be growing, and it is becoming increasingly important to the 5G business case. Above all, MEC can become a reality, particularly as more commercial 5G deployments begin to appear, services start to bed in, and yet more use cases start to emerge.