Listening to all the hype about LTE, you would be forgiven for thinking that once LTE networks are rolled out they will sweep away everything that went before, and we will all benefit from continuous fast mobile data connections. In fact, by 2016 LTE will still only account for 10% of the world’s subscribers, and coverage will be far from ubiquitous. The reality is that for many years LTE will need to co-exist in a complex environment with legacy 2G and 3G networks ‑ when leaving an area served by LTE, a data connection will need to fall back to HSPA, and a voice over LTE (VoLTE) connection would revert to UMTS or even GSM, or their CDMA equivalents. At the same time there is huge pressure on operators to maintain quality of service under all conditions, in order to keep their customers happy.
The concept of Self-Optimising Networks (SON) has only come to the fore with the advent of LTE. Typical use cases include self optimisation by adding or removing cell relationships and re-prioritization of relationships between neighbouring cells. This is called Automatic Neighbor Relationship (ANR). Another case is Self-Healing (SH), where the network responds to an outage alarm by changing the coverage footprint of neighbouring cells as well as making parameter changes to maintain coverage and capacity as far as possible. The third use case, automatic load balancing (ALB) alleviates network congestion by balancing between cells based on the load seen by each cell, using a combination of parameters as well as the cell footprint itself. Finally, Automatic Parameter Optimisation (APO) works by changing various parameters to improve network KPIs for existing cell sites as well as for new sites coming on air. It automatically tunes parameters and works towards achieving the required KPIs.
Although no SON specifications exist for 2G and 3G, it can also be very useful in optimising a multi-RAT network. Consider the scenario shown in Figure 1, where there is an outage in one of the LTE cells, which overlays cells on the older UMTS and GSM networks. By automatically initiating parameter tuning on the neighbouring LTE sectors, the network can adjust the threshold at which a user equipment (UE) will transition to the UMTS network. In this way it ensures that the handover occurs before the user enters the sector that is out, therefore avoiding loss of service while the UE hunts for an alternative network.
Real-life networks also utilize equipment from a number of different infrastructure vendors, so an inter-RAT handover will be a handover not only between different technologies but also between different vendors’ equipment. A SON offers the potential to optimize parameters so that this handover takes place as smoothly as possible. In this way a multi-technology SON, although not specified for 2G and 3G, has a far greater capability than a vendor SON that is unable to manage an inter-RAT transition to network sites utilizing another vendor’s equipment.