The terms edge computing and edge computing are commonly used interchangeably. However, the two have important distinctions. Edge Computing is a concept, and MEC is a standard architecture.
Within the broad topic of edge computing, MEC is the widely accepted standard that must be met for a technology to be considered edge computing. While not an industry mandate that products meet MEC standards to be billed as edge solutions, many vendors are building around the standard. The general term of edge computing covers the practice of offloading computing processes (and in some cases the handling of storage and networking resources) from the user’s computer or device to a local network node or other computer.
Edge Computing is an overarching term for the practice of placing compute and storage resources local to the end-user (at the “edge” of the network), to augment a distant central cloud. The edge can exist in places such as on the customer’s premises or the base of cell towers. In addition to traditional server racks, edge computing can take place on smaller pieces of hardware like routers or WiFi hotspots where practical.
Not sending data to a distant central data center for computing power means faster network speeds, lower latencies, and more reliable network connections.
Edge computing is a more general concept than MEC and less general than fog computing. Source: SDxCentral
Standards by ETSI
ETSI created a catalog of over a dozen standards and papers for various aspects of edge computing, or to use the term they created and use, MEC.
ETSI’s MEC standards are guided by the following principles:
- Edge technology should have a virtualization platform to be considered MEC (ETSI uses their NFV architecture in the standard).
- MEC can be deployed at radio nodes, aggregation points, and the edge of the core network.
- APIs in a MEC environment should be simple, controllable, and if possible, reusable for other tasks.
- Since the compute, storage, and network resources that a MEC application requires may not match what are available at a node, a MEC network needs a system-wide lifecycle management of applications to handle these variables correctly.
- MEC systems must be able to relocate a mobile edge application running in an external cloud to a MEC host and back while fulfilling all of the application’s requirements (ETSI admits this principle is in need of further study).
For more detail on these principles and MEC’s generic technical requirements, check out the paper here.
Note that ETSI previously referred to MEC as Edge Computing. Over time, the field expanded the focus of edge computing beyond mobile implementations.