Edge Computing and Routing

Edge Computing and Routing

Gartner predicts that “edge computing will become a dominant factor across virtually all industries and use cases as the edge becomes empowered with more sophisticated and specialized compute resources and more data storage.” The main driver for this focus on the edge comes from the need for Internet of Things (IoT) systems to deliver disconnected or distributed capabilities into the embedded IoT world.

IoT is one of the business applications enabled by 5G communication services. It shows why new 5G applications and opportunities will focus more on enterprises connecting things than end users connecting devices. For example, URLLC enables industrial, medical, drone and transportation applications where reliability and latency requirements surpass bandwidth needs. Categories for these real time applications include:

  • Connected and automated manufacturing and logistics operations
  • Smart City intelligent transportation systems (ITS) for smart traffic management to handle vehicle congestion and route emergency vehicles
  • Autonomous vehicles which will allow vehicle manufacturers to offer new services
  • AR (Augmented Reality) / VR (Virtual Reality)
  • Tele healthcare

These applications generate significant volumes of data transfer and data processing, which will require some form of cloud computing. Expensive computations cannot run on a mobile device due to its limited resources. However, for applications that require low latency, that computation cannot occur on conventional cloud data centers because they are concentrated in a few areas that are often quite distant from the user. Since signals cannot travel faster than the speed of light, distance translates into latency. These services will only work if the computing resources are much closer to the end-user or devices. This suggests that computing resources will need to be both still in the cloud and as close as possible to the end-user. The only possible place is at the edge of the network.

Network operators are ideally suited to leverage the opportunity of providing edge computing services.  They often have real estate within the network edge that can be used to house computing and storage. For example, central offices have the virtues of space, power, network connectivity, and proximity to the end users. This can be combined with existing data centers or colocation locations as needed.

From a market perspective, the research firm IHS Markit envisions a range of data center options to make computing resources available as close to the customer as needed, with MNOs focusing on Access and Edge data centers:

Multi-Access Edge Computing (MEC)

On one side, edge computing places high-performance compute, storage and network resources as close as possible to end-users and devices, at the edge of the network, to fulfill the real-time requirements of 5G-enabled business applications

On the other side, the disaggregation and virtualization of networking technology via network functions virtualization (NFV) require 5G operators to provide computing resources closer to cell towers and the edge to run their virtualized network equipment using concepts like Central Office Rearchitected as Datacenter in order to provide the necessary compute and storage resources to run NFVs. Therefore, the same edge facilities and resources can be used to support both business edge applications, and MNOs’ own NVF.

Given this rationale for the tight integration between edge computing and 5G, and the need for interoperability, standards are important. The edge computing standards from ETSI is Multi-Access Edge Computing (MEC). It is an open framework for applications and services that are tightly coupled with the Radio Access Network (RAN) via open interfaces to integrate software services into wireless networks.

Routing, 5G and MEC

5G transport networks must support the following requirements:

  • A disaggregated approach using virtualization for software-based network functions
  • Much more bandwidth per cell site
  • Many more cell sites, especially small cells
  • All IP over Ethernet down to the cell site
  • New services with stringent delay and bandwidth requirements

Since the networks will be IP based and there are stringent requirements including latency, service providers expect to use routing extensively. A well-documented example is the Telecom Infra Project’s (TIP) Disaggregated Cell Site Gateway (DCSG) project. TIP specifies the hardware and software requirements to connect a small number of cell sites per router. Moreover, they envision using the DCSG to connect and cover business and broadband services. The DCSG is expected to support MPLS, Segment Routing, QoS, and timing and synchronization. The ability to deliver different service levels is critical to services using network slicing. Fundamentally, using routing for the service intelligence in this type of platform will be critical to ensure the performance of the various applications running on the 5G transport network.

Since edge computing will be integral to 5G opportunities, routing is essential to ensuring service levels. The synergies of 5G and edge computing, particularly MEC, are clear.  At the simplest level, the compute infrastructure for NFV can also be used for edge computing services. To fully realize the 5G opportunities, edge computing will be required.