With multiple deployments already underway around the world, 5G technology is ramping up -- unlocking new applications with high-speed connectivity and near real-time latency for end-users, both humans, and machines. Due to the vast potential 5G provides, 5G connections are slated to reach 1.4 billion by 2025.
In comparison with existing 4G networks, 5G is expected to offer orders of magnitude improvements in download speeds, latency, and the number of connected devices. Of course, what we will actually experience will depend on a variety of factors, such as the performance offered by your particular service provider and where you are physically located when you access the 5G network, as location dictates coverage and performance.
As mobile network operators (MNOs) and wholesale operators alike update their network architecture to support 5G, it is important to note 5G networks are built upon technology that will disrupt the standard operating procedures of the wireline networks behind cellular connectivity, as airborne wireless bits eventually traverse expansive wireline networks.
Orders of magnitude increases in scaling existing wireline networks to support 5G is simply not economically feasible. Thus, it is essential to adopt automation, analytics, and agility to more intelligently and flexibly allocate network resources when and where necessary, for such things as network slicing that will offer guaranteed end-to-end network performance.
Impact of 5G on transport networks
MNOs should note the transition from 4G to 5G networks is not as straightforward as simply implementing a network capacity upgrade. Successfully enabling new 5G applications and business models will require a change in the way networks are designed, deployed, and managed. There is a need for enhanced xHaul (fronthaul, midhaul, and backhaul) transport networks to support the new end-to-end performance expected of 5G mobile services.
While today’s 4G networks rely on macro cell towers to carry the bulk of mobile traffic, 5G networks will leverage a much denser radio access network (RAN), composed of 10X to 20X more small cells, and therefore, 10X to 20X as many routers to interconnect them.
The denser network will also result in a significant increase in the associated operating expenses (opex) to manage this new wave of required routers. For that reason, existing network architectures are not economically viable, and unnecessarily complex, for network operators to rapidly and successfully deploy so many new small cells. This is where automation, orchestration, and software-based control come into play.
Via automation and orchestration, network operators can optimize resources while improving consistency in deploying network services. Automating quickly, consistently, and reliably configures network elements while eliminating human errors associated with manual change management. Orchestration is the execution of the operational processes involved in designing, creating, coordinating, and delivering end-to-end network services.
While IP technology is the industry choice for emerging applications such as 5G, the way IP-related protocols are delivered must evolve in tandem with network advancements. In fact, 5G applications are driving the need for more IP routing capabilities much closer to the network edge, closer to the RAN. This calls for a move to networks that are adaptive and agile enough to meet the changing needs of new-age technologies.
A different approach with Adaptive IP
Between the insatiable desire for more bandwidth, scaling, and flexibility in network deployments, an entirely new approach to network architecture has emerged that addresses these new network requirements - Adaptive IP.
Adaptive IP offers essential standards-based IP capabilities required to support new applications while enabling network operators to benefit from cloud-like scale, disaggregated functionality, and intelligent automat driven by real-time streaming telemetry. It supports existing services and applications while setting the stage for new and emerging use cases -- such as 5G -- by going beyond the individual capabilities of a router by taking a holistic view to solve networkwide challenges, not just standalone router platform challenges.
Bringing together software-defined networking (SDN) based orchestration with telemetry-driven analytics and automation, Adaptive IP enables network operators to turn up only the essential IP capabilities and take advantage of real-time network insights to automate services and facilitate operations. This results in a simpler, more streamlined IP connectivity that leads to economies of scale, made possible through a simpler, more open, and highly programmable network.
Software control and automation
Integral to adaptive networking are multi-domain service orchestration (MDSO) and centralized, software-defined control of individual domains. Through the implementation of SDN, network operators can simplify end-to-end management and automation of services across multi-vendor, multi-layer hybrid networks. This open, vendor-agnostic approach enables rapid creation, deployment, and automation of end-to-end service delivery, across both physical and virtual domains, and across both wireless and wireline domains.
New applications require compute power to be located at the edge of the network, delivering scale and performance dynamically as needed while significantly increasing the number of deployed IP-enabled network end-points. This associated complexity makes intelligent automation a key network requirement.
In any legacy IP architecture, an ever-increasing number of costly hardware-based router platforms perform this task. In contrast, Adaptive IP’s centralized software control and automation layer will reduce node-based control plane signaling, making deployments far simpler and more cost-effective to manage and maintain. This capability makes it easy for any network operator to start their unique journey from a box-centric, legacy IP approach to a simple, automated network design -- efficiently supporting legacy services while preparing for the next wave of application requirements. IP vendor lock-in is also eliminated.
Analytics and intelligence
Adaptive IP technology enables network operators to capture information about network topology, latency, routing, and other metrics, to create a simplified and unified virtual IP network map. This map provides a streamlined, near real-time view into how routing behavior is affecting service delivery and acts as a path computation engine (PCE) to determine what IP network parts need to be optimized. This information is also passed through open APIs to automatically configure service and traffic flows for optimal IP network performance, offering closed-loop automation that delivers optimal IP connectivity, from access to metro networks.
With the collection and analysis of network performance using machine learning, MNOs can proactively predict potential network issues and anticipate trends more accurately by turning streaming telemetry data into intelligent, actionable insights. Built upon an open and extensible microservices-based architecture, the robust and flexible framework for collecting, processing, and storing network data facilitates innovative analytics-based applications.
Leveraging these insights will help network providers develop smarter, data-driven business policies and enable them to adapt to customer needs securely and in real-time. This visibility also plays a key role in optimizing performance on an ongoing basis, such as by correcting service delivery issues and delegating resources more cost-effectively.
Programmable infrastructure
The Adaptive IP approach leverages both packet and optical technologies that are designed to securely manage a dynamic pool of virtual and physical resources so they can be accessed and configured via common, open interfaces. Leveraging a service-aware operating system (SAOS), network operators can evolve toward a disaggregated architecture, offering the capabilities required to progress the network by leveraging a microservices-based architecture and container-based software. The result is rapid service creation, deployment, and modification to maintain pace with user demands.
Operators will require a purpose-built infrastructure to evolve towards an open, flexible, and scalable Adaptive IP network. Such an infrastructure eliminates significant cost and complexity, as it is designed for specific applications using coherent optics targeted specifically at access aggregation and metro networks.
The future of IP networking is Adaptive
Adaptive IP leverages highly instrumented infrastructure offering the ability to export real-time network performance data used to meet the demands of applications running on top of it, on an ongoing basis. Adaptive IP will be instrumental in helping network operators transition to 5G. This approach offers optimization by automation, programmable infrastructure, and telemetry-driven analytics that together deliver the agility needed to allocate network resources with maximum return and performance.
Brian Lavallée is senior director, portfolio solutions at Ciena.