They’re struggling because the vast majority 4G transport infrastructures were designed to support consumer services with stable north-south traffic patterns, optimized for best-effort mobile broadband, and no need or ability to support stringent SLAs. Very few adopted advanced 4G features that would support more complex services, such as Enhanced Inter-Cell Interference Coordination (eICIC), Coordinated Multi-Point (CoMP), Evolved Multimedia Broadcast Multicast Services (eMBMS), Multiple-Input Multiple-Output (MIMO), or precise timing.
The expansion to 5G has changed the game, as three groups of services emerge – eMBB, URLLC, and massive machine-type communications (mMTC) – all driving a new set of requirements over the transport network. To match the commercial-grade SLAs these new services will inevitably demand, mobile networks must increase availability, reliability, and security.
To meet the service demands of 5G, service providers need to rethink the economics of their end-to-end infrastructure and make decisions that go beyond the radio. At the top of this list is the RAN architecture, which accounts for nearly 80% of service provider’s Capital Expenditures (CapEx). RAN operating costs are no slouch either; it’s estimated they equate to 60 % of the Total Cost of Ownership (TCO).
New densification strategies are also required to improve the economics of backhaul networks, supporting the high frequency (3,5Ghz, 6Ghz, 24GHz, etc.) spectrum with up to 400Mhz ultra-high cell bandwidth, massive MIMO (64TRX, 32TRX Multi-Access Antennas) and the exponential increase in base stations required to support 4G LTE expansions and 5G services.
Open, software-defined, virtualized, and cloud scalable, the emerging Cloud-RAN or Centralized RAN (C-RAN) architectures, support these densification strategies. C-RAN disaggregates base stations by separating Radio Units (RUs) from Radio Equipment Controllers (REC) and centrally locates these REC functions in regional/distributed data centers.
The centralization of base stations drives the need for fronthaul transport to be able to carry the antenna samples using CPRI or standard-based/open protocols. However, CPRI line rates (currently limited to 24 Gbps) don’t efficiently scale to meet the growing needs of 5G and also require tight delay budgets that lead to limitations in the distance and transport technologies that can be used for aggregation. Another concern is related to MIMO scale adoption as CPRI requires a dedicated link for every antenna. This mechanism becomes problematic as service providers invest in MIMO technology to increase data rates for 4G and 5G radio.
With these new architectures, service provider RAN departments must have a stake in ensuring the network between the RU and the DU delivers the required performance. This was not the case in the past as the RU and the DU were tightly integrated at the cell site.
These attributes are essential to the transport infrastructure meeting performance expectations and assuring a smooth evolution to C-RAN while supporting legacy RAN requirements.
Converged end-to-end IP infrastructure
To improve economics, RAN transport networks need to be optimized to share connectivity for both wireline, wireless and business services. This can be done effectively by leveraging technologies such as Segment Routing to address the diverse 5G transport requirements – for example, different SLAs can be enforced thanks to network slicing capabilities powered by Segment Routing directly from the cell site up to metro and core networks.
As operators move to Cloud RAN, mobile network functions are virtualized and distributed closer to end-users for better service quality and lower latency. Cloud RAN locations, in turn, need to support the dynamic placement of VNFs and Service Edge, while providing telco class resiliency; this is driving service providers to roll out multiservice networking devices that support end-to-end network slicing.
Designed to Evolve
As 4G and 5G RAN specifications and standards evolve towards cloud-scale, open, and virtualized solutions, RAN transport must be extensible to meet current and future RAN functions and technical requirements. Turbo-charging RAN networking devices with Field-Programmable Gate Array (FPGA) will increase the flexibility to quickly adapt to evolving radio interface processing specifications for CPRI, eCPRI and RoE.
Automated
To curb OpEx costs and better manage operational life-cycle, RAN transport solutions need extensive and open automation capabilities that can integrate into existing management domains and end-to-end systems. As densification can require on-site interventions, network operators need greater automation in deployment processes and tighter security requirements.
Packet-based
Traditional fronthaul optical solutions have been static TDM solutions using dedicated optical resources to carry fronthaul traffic. By contrast, packet-based solutions leverage statistical multiplexing to share transmission capacity, using valuable fiber resources more efficiently. Packet-based solutions also support an enhanced CPRI (eCPRI) protocol that scales bandwidth 10x more effectively than 4G CPRI, meaning fewer transport resources are required.
Packet-based solutions support Point-to-Point and Ring topologies, offering easy bandwidth growth while providing the telco class resiliency demanded by mission-critical enterprise services. Fronthaul optical transport solutions are typically implemented using point to point topologies.
Open
Already, service providers have shown an unprecedented interest in open, virtualized RAN solutions that dovetail into their broader SDN architectures. As an early leader in the creation of the Open vRAN ecosystem, we’ve witnessed great progress, but there still is a long road ahead. Continuing to grow and mature the Open vRAN ecosystem will ensure new open RAN solutions find their way into service providers’ infrastructures.