Reflecting the natural evolution of mobile Base Stations and the Radio Access Network (RAN), Multi-access Edge Computing (MEC) first emerged as an ETSI Industry Specifications Group (ISG) in early 2016. Originally defined as Mobile Edge Computing, the descriptor was updated in September, of that year, to reflect the fact that the same concepts described in the early requirements were relevant to fixed wireless, Wi-Fi and wireline access. The evolving ETSI MEC ISG documents can be found here.
Multi-Access Edge Computing infrastructures allow the implementation of software-only mobile functions or Software-as-a-Service (SaaS) applications that operate entirely within a standardized virtualization platform which is deployed in or close to the network edge.
Multi-Service Edge Computing reference architecture, per ETSI GS MEC 003 v1.1.1
The Multi-Access Edge Computing reference architecture consists of two functional areas -- Host and Management -- with the management layer comprising both host and system-level administrative entities. The combination of these functional elements provides the foundation required to operate a distributed environment for instantiating and scaling mobile applications and services in a highly granular and dynamic manner.
At the core of the mobile edge host is the virtualization infrastructure, which supplies compute, storage and network resources to the mobile edge applications. Data plane functionality, within the virtualization infrastructure, applies rules and enforces access control lists while routing traffic between network services and mobile applications. These may be local to the host or external on hosts residing within other networks.
The mobile edge platform, within the host, provides service and application assistance for mobile applications. This includes the authentication and authorization of mobile edge services, along with their discovery, advertisement and notification of any state changes. Under the direction of the mobile edge platform manager, the mobile edge platform controls the spin-up, quiesce and graceful termination of mobile edge applications. All data related to each mobile edge application instance is stored within the platform, including any transport dependencies, traffic rules and DNS rules.
Mobile Edge Application Start up
The mobile edge platform manager is responsible for the entire application lifecycle and informs the mobile edge system of any relevant application-level events. It also receives and processes status reports and performance measurement information from the virtualization infrastructure manager (VIM). The VIM is a critical component of ETSI NFV ISG architecture specifications, being responsible for allocating, managing and releasing virtualized compute, storage and networking resources -- generally preparing the infrastructure to run a software image.
A core element within the overlying mobile edge system, the mobile edge orchestrator maintains a holistic view of the mobile edge hosts, including their available resources and services. The mobile edge orchestrator on-boards and records applications, performing integrity checks while confirming the rules and requirements, for running the software packages, comply with operator policies. With a complete topological view of the entire mobile edge system, the mobile edge orchestrator can select the most appropriate mobile edge host from which to instantiate an application. This is generally based on the available services, required resources and infrastructure demands, such as latency.
External operations support systems (OSS) interface with the mobile edge orchestrator over mobile edge management reference point one (Mm1), the requirements for which are detailed within ETSI GS MEC 010-2. Additional Mm reference points are defined between orchestration elements, management functions and host platforms, while two Mx reference points interface with external entities. Mobile edge platform (Mp) reference points, to local or remote applications, provide service registration, discovery and communications interfaces. This reference point group also interfaces with the data plane, instructing the underlying infrastructure on how to route traffic between applications and network services.