The European Telecommunications Standards Institute (ETSI) says that multi-access edge computing (MEC) offers applications and content providers cloud-computing capabilities and an IT service environment at the edge of the mobile network. Virtually every conversation about 5G connectivity includes mention of edge computing and network slicing. Let’s examine each of these concepts and see why they are so critical to enabling 5G use cases such as connected vehicles (CV).
In a nutshell, edge computing moves data processing power from centralized cloud-based data centers to devices located far from those data centers but near where the data is needed. Edge devices include many varieties, such as gateways connected to sensors or radio towers. In the 5G context, edge computing is usually discussed as MEC. Typically, subscriber applications are deployed in the service provider’s data center or as Internet applications. MEC changes this model to allow applications to run on general purpose processor (GPP)-based servers at the network edge, close to the evolved NodeB (eNB), or even on the edge device itself. Think of it as a mini data center that can be located right in the cellular base station, or in a car.
The advantage of MEC, especially in the context of CV, is that the compute happens closer to where the action happens—where the cars and users are. This enables the cost-effective delivery of data to where the cars need it. MEC supports fast analysis—close to the point of capture—of the multiple exabytes of data connected cars are expected to generate each month. Edge computing can also facilitate the ability of mobile network operators (MNOs) to spin up and spin down network functions in response to specific user demands.
At the core of MEC is the MEC server. According to an IEEE Software Defined Networking newsletter, “the MEC server provides computing resources, storage capacity and connectivity as traditional cloud infrastructure would. Additionally, it provides access to user traffic and radio network information that can be used by application providers to tailor their applications and services for enhanced user experience. It hosts software for real-time analytics and machine-intelligence.”
Obviously, not all applications need to be at the edge. The “big cloud” is still going to be needed. For example, real-time high-definition map updates require both high bandwidth and low latency—that’s where 5G excels. But there's also great value for the bigger cloud to provide background information, such as the facilities located nearby, traffic updates, and weather reports.
As with any innovative technology, there are challenges to overcome before MEC can be deployed on a global scale. For example, the industry must determine how to publish APIs and software development kits (SDKs) for use by application developers and develop new network management and orchestration frameworks.
So where does network slicing come in to play? Essentially, it is an approach to network virtualization that can change the economic equation of building a cellular network and being able to monetize the data traveling on that network. In big data centers, virtualized fixed networks are common, using overlays, SDN and other techniques. 5G network slicing is similar, allowing MNOs to split a single physical mobile network into multiple virtual networks—essentially providing network-as-a-service capabilities that greatly enhance the network’s flexibility and responsiveness to user needs, and subsequently providing competitive advantage to those MNOs.
Network slicing can be used to partition the core network, but also can potentially be applied to various different radio access networks (RANs), or to different service types running across a single RAN. Each slice of the network can be customized to provide the appropriate latency or bandwidth capabilities for a particular use case such as delivery of video versus vehicle-to-anything (V2X) communication. While 5G network slicing is still in its infancy, commercially speaking, the Intel® Network Edge Virtualization (NEV) SDK is the first kit of its kind to provide an NFV platform targeted for MEC application and services.
To learn more about 5G, CV, edge computing and Intel’s role in helping build 5G networks that can augment the life experience of every ordinary consumer, watch the iGR webinar, “5G and End-to-End Enablement for Connected Vehicles and Autonomous Driving.”