5G: High Capacity, Always-On, and Ubiquitous

August 27, 2020 | Matt Anderson

Through the generations of cellular and mobile technology, the most apparent changes have been in the radio technology and new services enabled, whether digital modulation or text messaging. While every new generation is a leap forward in performance and new services, the overarching trend in the mobility infrastructure has paralleled the trend in other networks; all IP-based services, and all-IP-based infrastructure. The reason for this is simple: packet-oriented infrastructure matches the capability of the endpoints, the communication service provider/cloud infrastructure, and the software. In addition, all-IP networks are already proven to scale to large numbers of endpoints. 5G is no exception: IoT, vehicle-to-network, augmented reality, and ubiquitous computing will all require a pervasive set of interconnected IP networks ready to provide the flexible underlay infrastructure that 5G demands. 5G is more than a new radio; it is the culmination of mobility and Internet architectures converging with one another to create a network platform that is high-capacity, always-on, and ubiquitous.

While current networks can support 5G deployments in “standalone mode,” to realize the full potential the following abilities must be intrinsic in the underlay network:

Network Slicing: Slicing is the ability to build end-to-end segmentation with overlays for MVNO, hosted MNO, and private LTE/5G.  This includes slicing over a back/mid/front-haul provider network, as well as slicing for service tiers such as emergency first responders, latency-sensitive applications, and IoT devices. Segment Routing and SRv6 are the latest tools available to build segmented networks at scale.

Service Assurance: Particularly in a front-haul network, 5G has very strict performance requirements for latency, jitter, and loss. These enable various radio-over-packet approaches to enhance the performance of the RAN for network availability, mobility, and efficient use of spectrum. Deploying routers and switches deeper in the network to minimize the tower-to-tower latency is a necessary but not a sufficient solution: the network should provide real-time telemetry to identify service level problems before they result in an outage.

Rapid L3 Convergence at Scale
: When operating a highly-scaled network, outages due to fiber cuts, equipment failure, and utility problems have to be autonomically solved: the network must immediately recover on its own. High-performance routers and switches make this possible by:
Quickly detecting failures via hardware-accelerated Bidirectional Forwarding Detection (BFD)

  • Rapidly computing the new network topology in control plane software (BGP, IS-IS, OSPF)
  • Efficiently communicating the new topology to the rest of the network
  • Rapidly updating the forwarding state in the hardware (FIB programming)

If performance is lacking in any one of these steps, the network’s ability to support 5G can be impaired.

Edge Routing
: Since 5G is an Internet network, it will be deployed on an evolution of existing Internet edge routing deployments. This means that the edge gateway functionality will enjoy the same massive-scale 100GE and 400GE interconnect capacities, with massive bandwidth available for 5G services. It also means that security, access control, and infrastructure protection capabilities must exist in the edge router, including packet sampling (e.g. sFlow), DDoS and other policy steering (BGP Flowspec), and control plane protection (CoPP).

5G won’t be just the mobile instantiation of the Internet … 5G will be the Internet.  It will have the bandwidth, ubiquity, and service-readiness needed for always-on wireless access, to such a degree that it will end up surpassing many existing forms of wired Internet access and cloud services.  Recognizing the critical importance of this latest wave of network convergence will prepare investors, operators, and consumers to benefit from 5G as it is defined, marketed, and deployed globally.