Mobile Technology
The case for optimised silicon in the next generation of cellular networks
5G Open RAN Small Cells will improve security and flexibility and speed up deployment of a new era of network technology, explains Vicky Messer, VP product management, Picocom
W
ith the advent of 5G, networks are evolving to be open. Mobile Network Operators (MNOs) and
Communications Service Providers (CSPs) are looking to cut costs and speed up deployments to meet demanding coverage and capacity targets over the next few years. These low-cost and energy-efficient solution targets set by the operators make it challenging for cellular equipment vendors to meet the requirements for 4G, 5G, dual mode, macrocell and indoor and outdoor small cell infrastructure equipment. Equipment vendors are therefore looking to silicon vendors for flexible, cost- effective, power-optimised silicon with integrated features such as Open RAN and standardised industry interfaces.
5G Open RAN and disaggregated networks
Over the last 3-4 years, Open RAN has become a buzzword used by industry bodies and governments as the solution for diversification of the cellular network supply chain, allowing new entrants to compete with the dwindling number of incumbents. Operators’ cost pressures, need for network security and a desire for vendor diversification is driving vendors to demand stringent requirements and best practices.
In addition to the 3GPP-defined interfaces between the core network and RAN equipment, the SCF (Small Cell Forum) defined FAPI interface at MAC- PHY, new operator-lead organisations such as O-RAN Alliance have been developing and integrating open fronthaul interface standards for disaggregated networks. This ‘split’ in the RAN architecture includes the CUs (Central Unit), DUs (Distributed Unit) and RUs (Radio Units), enabling RAN equipment from multiple vendors to
40 February 2023
interface, increasing competition and driving down cost. This architecture provides additional capability to scale the network to meet required capacity and coverage.
Disaggregated networks, a concept conceived in the LTE (4G) days, using Cloud RAN solutions as a way to split the network, became popular, where some of the RAN processing could be offloaded onto COTS (Commercial off-
the-shelf) or X86-based processors and then interface to low-cost radio units. A CPRI interface was used between very low-cost radio units and the baseband unit (BBU); however, the CPRI interface needed standardising to allow for the diversity of vendors.
Another factor in the architecture of fronthaul or backhaul infrastructure is that new enterprise developments and green field deployments benefit from the investment of fibre between the core network, baseband and radio equipment. This enables high transport bandwidth and low latency fronthaul interface between the DU and RU, allowing more RAN processing in the DU and a simpler RU. For deployments where fronthaul infrastructure is ‘non-ideal’ copper, fully integrated RAN, or splits 2, 6 where more processing is at the radio unit is preferable. Deployment use cases also come into play with the RAN architecture and specification, which include: ● Large outdoor base stations providing wide-area coverage tend to have mMIMO (massive MIMO) with as many as 64 antennas to provide high efficiency with directed beams. ● Outdoor microcells utilising low band spectrum with up to 4T4R or 8T8R wide
Components in Electronics
area for coverage holes, demanding power efficient solutions. ● Indoor enterprises will have limited power delivered to the radio unit due to PoE (power-over-Ethernet) classes and heatsink-only power dissipation for wall or ceiling mounted radio units. ● Neutral host vendors also demand multi-carrier solutions within the radio unit, requiring 2-3 times wider bandwidth than the 100MHz FR1 bandwidth. ● The diverse worldwide 5G NR spectrum serving the use cases is even more significant than the challenge of LTE, with a wide range of requirements for frequency channel, bandwidth and carrier aggregation support resulting in customer-specific RF front-end designs and flexible baseband solutions. Yet, customers are demanding reference designs from silicon vendors. ● Low latency applications, such as video on demand, require Edge servers for maximum performance and efficiency. ● Dual mode 5G/LTE radio units, NSA (Non-Standalone Access) and renewal programs demand concurrent 5G and LTE support within the same radio unit whilst minimising the BOM cost.
5G presents challenges, Open RAN offers solutions New 5G networks present a myriad of challenges to vendors, such as ensuring the implementation of low-power solutions that are critical to the implementation and success of 5G due to environmental restrictions and fuel cost economics to meet sustainability targets for all infrastructure deployments. In addition, radio units operate without active cooling dissipation and, in many cases, are required to function within PoE power constraints. Next-generation open interfaces, defined by 3GPP, O-RAN Alliance and SCF, are further optimising and developing standards such as the O-RAN Alliance Open Fronthaul. Mature optimised SoCs now include interfaces such as eCPRI. Flexible RAN performance is required to meet operator, private and neutral host customer needs and ensuring frequency spectrum requirements are met, including carrier bandwidth. But first-generation solutions are always non-ideal, and generally based around expensive and power-hungry servers, accelerator cards and FPGAs. While Open RAN is not a magic bullet, it can (and does) address several of these challenges and presents a quicker and more cost-efficient path to market for 5G vendors and operators alike. Operators are looking for lower BOM (bill of materials) cost targets at volume deployments with optimised silicon. The diversification of vendors within the Open RAN ecosystem fosters competition to drive down these costs. Silicon vendors such as Picocom have optimised baseband SoC products and a product roadmap to drive down power. Hardening functions such as DPD (Digital Predistortion) algorithms in baseband/RFIC also contribute to
www.cieonline.co.uk
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62
