Mobile transport networks are on the threshold of a major transformation driven by the demands of 5G wireless technology. The demands include greater coverage, connectivity, and availability, along with dramatically improved speeds and latency.
A key goal for mobile operators is to enhance revenue from traditional markets while developing revolutionary new services and applications that take advantage of the new 5G infrastructure. To do this, operators need to leverage existing investments while also deploying new radio access network (RAN) and transport architectures. These will leverage both WDM and packet optical solutions that meet the transport requirements of existing and future applications.
Diversity of 5G applications and requirements
While 4G networks are primarily designed to carry a single type of traffic (namely mobile broadband), 5G networks are natively designed for multiple traffic types with very different requirements. 5G applications fall under three broad categories:
1. Extreme mobile broadband (eMBB) services offer higher bit rates and support extreme traffic densities for the evolution of communications and entertainment. For example, applications could include immersive virtual reality or 3D video applications where high capacity can boost throughput and provide reasonable performance everywhere.
2. Internet of things (IoT) services requirements are expected to drive 5G. For instance, massive machine type communications may soon connect billions of sensors, meters, and machines This will entail adding massive densities of low-traffic devices and bearers.
3. Mission-critical services and control will enable reliable, secure, and low-latency communications. Applications in this category include ultra-reliable and low-latency use cases, such as remote surgery. In this case, latency requirements could be as low as sub milliseconds or a few milliseconds.
5G-ready transport networks—what is needed?
Operators need a single “system of systems” capable of supporting 5G requirements without a new network build-out. This will require:
• High-capacity data rates (to support ten-fold increases in user data)
• Low-latency transport (to support real-time M2M applications)
• Distribution networks that can grow to support ultra-dense deployments (to enable cell densification)
• Ultra-high reliability capabilities (to support mission-critical applications)
• Highest-quality synchronization distribution capabilities (to support new RF interfaces)
Leveraging Ethernet for mobile transport
To leverage performance gains and use spectrum more efficiently, operators have been turning to C-RAN architectures that centralize the baseband processing. This has necessitated mobile fronthaul networks to deliver high capacity and low latency transport of CPRI or OBSAI protocols.
With 5G and the use of more antennas (e.g., massive MIMO) in addition to more spectrum, the capacity needed on transport links is daunting. That’s why there has been a significant push to leverage packet Ethernet networks. Key to realizing this is a functional split between the radio equipment and the radio equipment control, as well as a redesign of the fronthaul interface. The split which could be done at different levels permits the relaxation of the bandwidth or latency requirements. This enables the transport of next-generation, fronthaul interfaces (NGFI, eCPRI) over Ethernet fronthaul/midhaul networks. At the same time, though, this imposes new requirements on existing mobile transport networks:
1. Low latency/time-sensitive networking – Ethernet fronthaul networks will need to provide both low latency and high reliability. Time-sensitive, 5G-related applications and new sophisticated mobile devices will mean tougher requirements for Ethernet. This will require provisions for deterministic and low latency transmission, especially for mission-critical traffic. For this reason, open standards continue to be developed to ensure the reliable and timely delivery of Ethernet traffic. The IEEE 802.1 TSN standards span many projects and applications, including automotive, professional audio/video, and other industrial applications. A subset of these is now being defined for fronthaul as part of the 802.1CM standard.
2. Synchronization distribution – In traditional fronthaul networks, the frequency and phase sync are carried within the synchronous CPRI protocol. With the move to asynchronous packet-switched networks, such as Ethernet, a synchronization challenge arises. As a result, a means to maintain accurate synchronization must be established as packet-switched networks introduce variable delays. To overcome variable delays, which adversely impact RAN performance, the 1588 Precision Time Protocol can be implemented.
3. QoS/Burst tolerance – To assure QoS, transport systems must have high tolerance for traffic bursts. Using large packet buffers for burst absorption avoids packet loss in the presence of bursty traffic, and can handle contention when traffic on multiple ingress ports competes for the same egress port.
4. High resiliency – A robust network will be needed to support mission-critical services. Protection mechanisms, such as Ethernet Ring Protection and MPLS-TP providing sub 50ms protection, will have a role to play. In addition, the introduction of new redundancy and failure detection capabilities are being looked by the IEEE as part of the 802.1CM standard.
Furthermore, mobile transport networks will need to provide enough flexibility and capacity to support a mix of distributed RAN, centralized RAN, and future cloud RAN deployments. This will demand a universal” anyhaul” mobile transport network capable of supporting all possible fronthaul, including CPRI, OBSAI, real-time and near-real time Ethernet, as well as IP/Ethernet backhaul options.
Towards programmable networks
The increasing diversity of applications and requirements has created the need for “network slicing.” Network slicing enables operators to combine virtualization with their physical infrastructure in order to tailor network “slices” to the requirements of each application. Slicing requires a high degree of programmability, which can be provided by technologies, such as SDN and NFV. More fundamentally, slicing requires cloud-based baseband processing, as well as a virtualized, reconfigurable transport network capable of delivering low latency and high capacity.
A carrier SDN-enabled programmable fabric creates a flexible, scalable, and dynamically reconfigurable optical network. It supports access, metro, and core networks using physical and virtualized transport components controlled with smart SDN principles and Application Programming Interfaces (APIs). The smart fabric underpins the transformation towards next-generation mobile networks and provides secure, dynamically interconnected services. For innovative new service offerings, this ultimately reduces the time to market while enabling operators to gain first-mover competitive advantages.
