The war over the mobile network optimisation has come down to a scrap between centralisation and distributed intelligence. E&T draws the battle lines.
The rapid evolution of network technology means operators need to be smarter about how they manage their networks if they are to streamline their processes, cut their operating costs and increase their return on investment.Fixing 3G networks is time-consuming and costly, due to the manual and fragmented processes that many operators use. The introduction of LTE (long-term evolution), often as an overlay to 2G and 3G networks, is making it worse. LTE networks, which are likely to use more base stations than previous-generation networks to achieve the capacity and coverage that users have come to expect, will be more costly unless a better approach is found. Self-organising network (SON) technologies and greater automation of the business processes involved in running a network may be the way forward.
SON should enable automation of dozens of common network planning, configuration and optimisation tasks that can take days or even weeks to resolve manually. The resulting efficiencies should cut costs and ensure that problems get fixed more quickly. However, although SON shows promise, it’s still unclear how the technology will be implemented and how this will affect future networks.
There are three major approaches to implementing SON: centralised SON (C-SON); distributed SON (D-SON); and hybrid SON (H-SON). Each is backed by different industry groups because each approach has different implications for the future of mobile networks.
C-SON systems will give operators transparency and control over their networks, even if their equipment comes from multiple vendors. This approach means operators will be able to buy their equipment from multiple sources and adapt standardised SON processes to their needs.
The D-SON approach means equipment vendors can offer differentiated equipment but are likely to use proprietary SON algorithms and operations and maintenance systems and so restrict the operator’s control.
The H-SON approach puts some of the work in the network nodes and some in a centralised management system, combining benefits and drawbacks of the other two approaches.
As LTE networks are built, operators will want to get the most out of their new investment as well as out of the 2G and 3G networks that they run. SON can help with the planning and optimisation of these networks, and their ongoing management. But the operators’ choice of SON approach will define how they go about this.
The vision of SON is that networks can be planned, optimised and maintained at the push of a button. In reality, this is a long way off, but telcos are taking their first steps towards fully automated environments.
With the standardisation of mobile network management processes in LTE, the automatic or even autonomous optimisation of network parameters is coming closer to reality. The SON concept began in 2006 as industry groups started thinking about LTE technology, and how they might overcome the limitations of 2G and 3G.
SON has three major parts: self-configuration, self-optimisation and self-healing.
In self-configuration mode, when a new LTE base station is installed and turned on, SON functions enable it to download the latest software release from the operations and maintenance system and automatically configure the transmission settings of the backhaul network. Once this is complete, the basestation can start working. Self-optimisation functions can then maintain and improve the performance and service quality, using performance indicators sent from mobile terminals. Finally, in its self-healing mode, the basestation can detect network failures and act to minimise the resultant loss of coverage or capacity.
Network optimisation is a closed‑loop process that involves choosing an initial set of operating parameters for the network, deploying them, measuring the performance, looking for possible optimisations, and deploying new parameters to achieve them. The data that drives the optimisation process can come from various sources, such as performance measurements from the operation and maintenance system.
Self-optimisation will help operators cut costs by reducing manual intervention. Moving to full automation is likely to take a number of steps, with human involvement shifting to a higher management level, supervising SON processes and designing policies for SON functions.
SON will also enable operators to adjust the performance of specific network functions, such as handover, to provide a better end-user experience and greater network capacity. Network handover can be optimised by automatically adapting cell parameters to adjust the handover boundaries, based on feedback from performance indicators.
Self-optimisation will also enable dynamic network optimisation to make it more robust to environmental changes. Optimisation can be driven by capacity, coverage or performance issues. Other strategies will help save energy, cut interference, improve the service for mobile users, balance loads between cells, or improve the coordination between cells.
As mobile telecoms becomes more competitive, operators and equipment providers are increasingly at odds about the balance between centralised and distributed SON. Operators want to buy LTE equipment from many vendors. Using open standards enables a variety of independent third-party tools to be applied to configuration, testing and troubleshooting. Equipment makers, on the other hand, want to deliver standardised SON functionality but implemented in proprietary algorithms embedded in the basestations.
The Next Generation Mobile Networks (NGMN) Alliance is championing the H-SON standard, with an emphasis on C-SON features that enable visibility and control of the distributed SON processes, while maintaining the flexibility for operators to adapt standardised SON processes.
C-SON is ideal for planning and visualising diagnostics in a heterogeneous, multi-provider and multi-technology environment. Meanwhile, standards for D-SON are being driven by the 3G Partnership Project (3GPP), which has produced standards that establish a SON environment more suited to the preferences of equipment makers – focusing SON on proprietary algorithms embedded in LTE basestations, potentially limiting operational freedom for telcos.
There is no debate that certain SON algorithms are best implemented in a distributed way. However, the lack of coverage in 3GPP standards of the planning and centralised optimisation requirements defined by NGMN has left a gap between the standards and the needs of operators.
C-SON addresses these needs by enabling a tight integration between an operator’s planning systems and its multi-vendor networks. C-SON also enables multi-vendor control of D-SON and rich visualisation, analysis and multi-technology optimisation capabilities to allow operators to adapt standardised SON processes to their needs.
With SON playing a critical role in LTE networks, the differences between the NGMN requirements and the 3GPP standards is causing debate over how the technology will be implemented. Some operators are questioning whether the 3GPP standards will meet their needs for transparency and openness, and how their need for C-SON will be addressed.
Why should we care about automating a 4G networking technology? The answer is that SON technology should offer key operational and financial benefits for operators thinking about investing in next-generation networks, reducing their total cost of ownership.
Optimisation systems are now being deployed for 2G and 3G networks as operators try to apply NGMN’s automation vision to cut the costs of running their existing networks. These optimisation systems use many of the principles for LTE C-SON, as laid out by NGMN, and so may offer a way for operators to move towards centralised, multi-technology SON capabilities that could also apply to future LTE implementations. This means that mobile operators should be able to deploy comprehensive SON systems that help them save money while maintaining the flexibility they will need in the era of the mobile Internet. *
Jeff Atkins is vice-president for product management at Actix