What are the myths and realities of 5G and where are network trends taking us?
VanillaPlus talks to Colin Bryce, director of International Business Development at CommScope about the lessons to learn in 5G network roll-outs.
VanillaPlus: What are the latest trends you have identified in network densification, virtualisation, optimisation and simplification?
Colin Bryce: I would say these trends are not directly related to 2018 but are ongoing trends in the market that we think will continue as we drive towards LTE Advanced and 5G deployments.
The first is Densification. If one looks back historically it is easy to see that while spectrum availability and new radio technologies have played an important part in adding capacity to wireless networks, the bulk of capacity gains have been delivered through network densification
s (indeed the very concept of “cellular” networks is nothing more than a technique to densify radio coverage). We do not see this trend changing, especially in major urban areas more cells will be deployed in any given geography, either in the macro layer or in underlay heterogeneous networks to deliver capacity. Indoor coverage and specialist coverage solutions (like stadium systems) can also be considered a part of a densification process.
However, just because cell sites are getting more densely packed does not mean that the fundamental deployment issues of site acquisition, site power and site backhaul are getting easier to solve. CommScope is working on a range of products and solutions to ease problems related to these issues. We are developing a family of site infrastructure solutions involving street poles with radio mounting options either at the top, middle or bottom of the poles that will make it easier to push suitable sites through planning approvals processes.
We are also developing power solutions to deliver power remotely to sites and to more efficiently power active radio units mounted at tower-top. Finally, CommScope is a leading player in the provision of optical fibre deployment deeper into telecommunications networks and these deep fibre roll-outs can support dense site connectivity as part of converged networks.
Secondly, Virtualisation is the process by which the functions of a telecommunications networks can be enabled in software that runs on standard computing platforms or blades, rather than specially designed and customised hardware. The intention is that this will bring benefits to network operators in terms of costs (both capex and opex) but perhaps more importantly give far greater flexibility to quickly develop new service propositions for end-users and allow more efficient network utilisation by deploying functions at network nodes that minimise traffic distribution.
Virtualisation has been underway in many operator core networks for some time and it is probably true that it will have the biggest overall impact for network operators in larger core network nodes. The process of virtualisation has been slower to develop in the Radio Access (RAN) portion of the networks due to perhaps the more widespread usage of specialist ASIC and DSP processors needed for real-time functions specific to each type of deployed radio technology.
Following Moore’s law, however, the speeds and capabilities of families of general processors are now able to deliver much of the digital processing requirements of radio base stations. As a result, we are starting to see the initial stages of virtualisation taking place in the RAN.
Standards bodies (3GPP) have now defined the functions of a base station for 5G into three parts:
- Central Unit (CU): Likely to be a cloud-based component implemented on COTS hardware.
- Distributed Unit (DU): Likely to be a combination of COTS and Customised hardware. Handles real-time, Layer 2 and 1 functionality
- Radio Unit (RU): Delivers RF functionality and physical layer connectivity.
The exact split of connectivity functionality between the DU and RU is being discussed and will offer a few options that will trade distributed functionality for point-to-point capacity reduction. As well as offering the previously stated advantages of virtualisation, defining the base station architecture in this way means it will be easier to deploy features that require radio coordination across multiple radio sites. Aggregation points for CU or DU only functionality will also be obvious places to deploy hardware to deliver Multi-access Edge Computing (MEC) capability.
The next trend is Optimisation. In both LTE and 5G networks the delivery of higher data rates depends not just on more bandwidth and carrier aggregation but also on utilising higher modulation rates like 256-QAM and 8T/8R architectures for non M-MIMO deployment). To make maximum use of these technologies the system needs to deliver very good Signal to Interference and Noise (SINR), for example MIMO needs SINR of around 13dB or better before it is likely to deliver any significant network benefit. Similarly, modulation methods will drop back to 64QAM or worse if the underlying network quality is poor.
Well -designed radio products and a good radio plan are pre-requisites to good SINR across networks. However, if these are not correctly implemented in the network the result will be sub-optimal. Field experience has demonstrated that a significant proportion of sites (often depending on country and network operator but none are immune) fail to meet the highest standards.
Issues like, incorrect antenna boresight settings, incorrect mechanical or RET tilt setting, crossed cables, and poorly made connections leading to PIM and feeder reflections are prime causes of network performance degradation. It is thus very important that operators continue or develop programs of site audit and issue-rectification and have well documented site acceptance criteria after new install or upgrade work is completed.
Finally, there is Simplification. In some ways, this can be considered a slight oxymoron, in the sense that as technologies evolve from analogue 1G through to 5G, the underlying networks’ principles and technology have undoubtedly got more complex. At the same time, however, it has become more expensive and harder to find well qualified technician level engineers in most markets and number of sites deployed has grown rapidly. For this reason, simplification is really about levels of abstraction. In other words, hiding the underlying complexity behind easy to use interfaces or making the tasks necessary to deploy, test and upgrade network equipment simpler, easier and more fool-proof.
At a network level this can be understood through initiatives like Virtualisation and Self-Optimisation but even at the physical layer that CommScope focuses on, we are constantly trying to improve usability by initiatives such as; Smart DC by-pass filters, pre-connectorisation on cables and fibre and incorporation of sensors that feedback data to central monitoring locations.
VanillaPlus: How will LTE continue to be foundational and how will it underpin the 5G network of networks?
Colin Bryce: I believe Network Operators face two main (and in some ways competing) challenges over the next few years:
The first is delivering enhanced performance on 4G-LTE networks:
- Enabled by the evolution of LTE standards (LTE Advanced and Advanced Pro)
- Delivery of LTE in unlicensed spectrum
- Support for Cat 16 devices
- Re-farming current 2G/3G spectrum and enabling greater carrier aggregation
The second is deploying 5G-NR onto new spectrum allocations. You need to:
- Understand the business case and develop appropriate deployment strategies
- Acquire the optimal spectrum allocation
- Make the correct economic/technology decisions, and
- Master new customer positioning.
Understanding the relative timing and scale of these two activities and assigning appropriate budget and resource to each task will be critical to the overall commercial performance and competitiveness of individual operators. CommScope believes that in most markets it will be possible to deliver 1 Gbps LTE by selecting the correct total bandwidth, modulation and coding schemes and utilising MIMO. Doing this will allow operators to undertake benchmark marketing for enhanced new features without having to very quickly deploy wide-scale 5G network capability.
It should also be noted that initial 5G will be deployed in the “non-Stand-Alone” (NSA) mode and will require an LTE network to be in place to provide the control-plane signalling to the end user devices. It is likely that initial 5G roll-out will not be like previous generations of technology in that it will not be rolled out on a national geographical scale, but will be deployed only in urban areas with high capacity needs and in localised industrial areas where new services can be markets tested.
The bulk of enhanced broadband upgrades to average users will be delivered using LTE-Advanced and LTE-Advanced-Pro. Getting the 5G implementation plan right in terms of balancing the timing of coverage and growth in real revenue generating service against the need for marketing impact by showing commitment to the new technology is going to be an interesting challenge!
Jeremy Cowan was talking to Colin Bryce, director International Business Development at CommScope.