Mobile Network Optimisation Evolution with SON

Christopher Larmour, CMO, Actix, describes the importance of SON, Self Optimising Networks in today's mobile cellular telecommunications environment.

Cellular telecommunications operators are starting to invest heavily in LTE, with the aim of providing innovative services and a superior user experience to mobile phone subscribers. But for that to happen, they need to learn from their 3G experience and optimise networks, instead of focusing solely on expanding the cellular telecommunications infrastructure. Although mature, today's mobile networks still drop about one in every fifty calls - a real issue for the modern subscriber.

SON (Self Optimising Networks) provides a solution for operators planning to roll out LTE networks as it prevents and responds to common failures, optimisation and configuration issues that would otherwise be lengthy and costly to deal with in a manual way. SON cuts engineering and technical costs and often reacts at a quicker rate to problems which, if untreated could go as far as affecting the customer's experience.

Mobile network optimisation drivers

As LTE cellular networks go live, it is expected that 2G, 3G and LTE technologies will coexist for some time. However, the incremental cost of delivering LTE will have to be low because 2G and 3G have already been cost optimised and operators' EBITDA headroom is tight.

LTE brings with it new engineering parameters, which will require more effort to optimise and lead to instability in the network, particularly in early deployment phases. Before deployments go live, it is necessary to have self-optimising processes in place.

The demand for FMC services, combined with the need for cellular operators to dampen the drain on capacity created by mobile broadband has created a new market around femtocells. However, the volume and variety of eNodeBs will drive the need for plug and play aspects of SON functionality.

Crucially, both the 3GPP and NGMN Alliance have recognised the need for SON, seeing it as a key functionality in Next Generation Networks. As LTE is tried and rolled out throughout 2010, it will be imperative that opcos include SON networking components in their systems.

Several leading operators are strong supporters of SON. Kenny Graham, head of new technologies and innovation at Vodafone, highlighted the crucial role of SONs for homezone LTE deployments, while T-Mobile UK's director of technology, Emin Gurdenli, described network automation as critical to investment in next-generation wireless technologies.

Components of cellular SON systems

An up-to-date accurate view of RAN status is essential to efficiently self-optimise a network. Scalable, multivendor Network Status Management data platforms should be used to consolidate data from Performance, Fault, Configuration, Planning and Call trace systems.

Next, a system of rules-based diagnosis and prioritisation must run in the background monitoring every aspect of the network - constantly sampling and analysing live data to identify predefined events, e.g. a dropped call rate crossing a predefined threshold. Events are prioritised according to urgency and impact on the network while rules encapsulate the team's location-specific RF Engineering knowledge. Once a problem has been classified, a solution-finding component is triggered.

Actions must then be taken to affect network conditions. Here the solution finding capability can either be autonomous (no human intervention required) or automatic (some human intervention required). However, while SON use cases have been defined, the specific algorithms or solution finding functions still have not. Therefore experience developing Automatic Optimisation algorithms for use in the field are a critical component of a live SON system.

Finally, coordination of control signals and execution of network orders by a control module are necessary to ensure signals to the network do not conflict with each other.

With the exception of control signal coordination, these capabilities are not the preserve of base station infrastructure manufacturers. As counterintuitive as it may seem, these vendors have not been traditionally strong in these areas for two main reasons: firstly, infrastructure vendors have generally focused on manufacturing base station hardware, deploying and operating networks but not designing commercial grade RF software tools, which is considered a distraction to core business.

The second reason rests on their customers: Mobile operators desire to keep their suppliers in check with independent third party tools used for configuration, testing and troubleshooting. Therefore, general-purpose optimisation tools provided by infrastructure vendors have never been truly embraced by those mobile operators who have some level of bargaining power. These vendor/customer dynamics will be present when SON systems start coming to market this year.

SON lifecycle

SON has three modes of operation: Self Configuration, Self Optimisation and Self Healing. In Self Configuration, when a new NodeB is "plugged in" it downloads the latest software from the operations and maintenance system via the backhaul network. The NodeB then scans the radio environment and adjust its physical or soft parameters accordingly. Once self-configuration is complete, the NodeB enters the operational state. Here Self Optimisation functionality is to maintain and improve service quality and performance of the network, taking into account performance indicators sent by mobile terminals. Finally, in Self Healing mode, the NodeB detects network failures and takes remedial action to minimise coverage or capacity loss.

3GPP standard 32.500 Release 2008-12 defines three types of SON systems:

1. Centralised
2. Distributed
3. Hybrid

With Centralised SON, algorithms are executed in the Operations and Maintenance System (OAM) with the SON functionality residing in a small number of locations, at a high level in the architecture. At the other end of the scale is Distributed SON, where SON algorithms are executed at the Network Element level, with SON functionality residing in many locations at a relatively low level in the architecture. Hybrid SON is where part of the SON algorithms are executed in the OAM system, while others are executed at the Network Element level.

Balance of types of self optimising network configurations.

Considering the large number of cells typical of next generation deployments and the complexity of the multivendor environment, centralised SON architecture appears to be the most comprehensive to manage variability and scope. Mobile operators will also require openness and visibility into SON systems, which also favours centralised SON. However, the centralised SON architecture - while likely to deliver the full complement of SON functionality - is likely to be more costly than more distributed SON variants.

SON Use Cases

Operator expectations are high for what LTE and SON should deliver with many keen on a step change in production costs. To provide clarity to vendors, the NGMN alliance identified typical use cases. SON use cases are categorised into the following groups: Planning, Deployment, Optimisation and Maintenance.

Planning includes preparing the necessary parameters and settings required before deploying a new enhanced NodeB (eNB). Once planning is complete, the new eNB can be placed and attached to the network.

The deployment of a new eNB consists of installing the eNB hardware at the designated site, and setting up the necessary configurations preceding commercial operation. Although mainly a manual process, effort can be reduced by having self-configuring functionality in the new eNB. This may also entail relevant updates in neighbouring eNBs as well as associated aGWs.

Optimisation is a closed loop process of parameter deployment, performance evaluation, parameter optimisation, and redeployment of optimised parameters to the network. The optimisation algorithm depends on the operator policy, and would be implementation specific. It is however essential to identify all necessary "hooks" to realise self-optimisation, such as UE/eNB measurements and interfaces, and to standardise these hooks to cope with multivendor scenarios.

The radio optimisation process can be capacity, coverage, or performance-driven. Coverage-driven optimisation means parameters are changed to increase cell coverage, while capacity-driven means parameters are changed to increase the network or cell capacity. For performance-driven optimisation, parameters are changed to increase user performance through the throughput per user or user plane latency.

When it comes to maintenance, the aim of SON is to allow smarter operation and maintenance in LTE compared to the 2G and 3G Networks today. Many use cases are based on direct inputs from operations and planning departments and consequently these use cases are related to daily operations and the most annoying problems there.

Developing a SON system is not a small undertaking. It demands extensive experience in RAN deployment and automatic optimisation. However, even just understanding these component parts is not enough. A SON needs operational experience in processing terabytes of data per day; such is the massive scale of a typical tier 1 mobile network.

Solutions such as Actix' Network Status Management (NSM) lies at the heart of a SON system. Delivering rules-based diagnosis and prioritisation of trigger events, NSM processes and correlates network data enabling automatic diagnosis of root causes. Automatic prioritisation of each issue is based on KPI impact to the network such as calls or revenue. When combined with automatic optimisation such as automatic frequency planning or automated radio parameter planning, these two core building blocks enable operators to deploy SON systems based on mature and proven technology. During 2009, Actix deployed the world's first SON system, comprising a full network model that will understand the macro context across several vendors, which will be able to manage heavy loads in real time, integrating new use cases easily and quickly.

As the shape of LTE networks comes into view, operators will be focused on not only maximising the value of new network investments, but also out of their existing legacy 2G and 3G networks. By extending the NSM and automatic optimisation systems already deployed in their network to form part of a Self Optimising Network, mobile operators will be able to free their network to respond to the new business models demanded by mobile broadband services.

Chris Larmour is Chief Marketing Officer with Actix, a company that has developed intelligent, automated Network Performance Engineering (NPE) and Network Status Management (NSM) systems. Actix solutions, which include a full suite of Self Optimizing Network (SON) capabilities, feature embedded wireless expertise, automating key processes and enabling very significant efficiencies in the deployment. Actix has delivered systems to 227 operators - including all ten of the world's largest mobile networks - in 106 countries globally.