Thursday, 23 July 2009

On Self Organising Network Concept in Rel-8 and Rel-9

Self-Organising Networks popularly known as SON are feature of 3GPP Release 8 and Release 9. SON has been around for quite some time now and is not a new concept. Its not an evolutionary technology, rather a revolutionary technology. The first time I heard of SON was in relation to Femtocells. Remember, a Femtocell has to start in an unfamiliar environment, learn about its surrounding and then adapt to the environment.

Other terms often used to mean SON is 'Plug-n-play' or 'PnP', 'Zero Touch', 'Auto Configured', 'Self Managed...', etc. SON is a very useful feature that will allow for the automation of several tasks lowering the OPEX costs. Examples include plug and play or a cell in between existing ones, neighbour recognition and (re-)configuration, optimizations, etc. Properly implemented, it could kill off drive-testing.

In simplest of terms, SON can be explained with the basic diagram above. A new cell created in an existing environment possibly due to too many existing resources being in use or too many users in an area during a particular time (football match for example) and this cell has to look at the surrounding and adjust its conditions. The other existing cells also have to adjust tehmselves with the change in surroundings.

According to recent analysis in NEC Whitepaper on SON (available here), about 17 % of wireless operator’s CAPEX is spent on engineering and installation services. SON’s self-configuring functions are expected to eliminate many on-site operations for the basic settings and subsequent updating of network equipments, and thus reduce CAPEX.

It is also known that about 24 % of a typical wireless operator’s revenue goes to network OPEX, which are the cost of network operation and maintenance, training and support, power, transmission, and site rental. SON’s self-optimizing functions will reduce a workload for site survey and analysis of network performances, and thus reduce OPEX. Moreover, SON’s energy-saving functions reduce the costs of power consumed by the equipment.

Self-optimizing and self-healing architectures improve user perceived qualities by mitigating quality degradations that result from inaccuracies of the planning or equipment faults as early as possible and by optimizing the network parameters under interference and overload conditions.
Nomor research has got an excellent paper on SON with regards to LTE. The full paper is available here. Here is an extract from that paper.

The main functionality of SON includes: self-configuration, self-optimization and self-healing.

Self-configuration process is defined as the process where newly deployed nodes (eNBs) are configured by automatic installation procedures to get the necessary basic configuration for system operation

Self-optimization process is defined as the process where UE & eNB measurements and performance measurements are used to autotune the network

Self-healing function aims at automatic detection and localization of most of the failures and applies self-healing mechanisms to solve several failure classes, such as reducing the output power in case of temperature failure or automatic fallback to previous software version.

A Self-configuration Subsystem will be created in OAM to be responsible for the selfconfiguration of eNB. For self-optimisation functions, they can be located in OAM or eNB or both of them. So according to the location of optimisation algorithms, SON can be divided into three classes: Centralised SON, Distributed SON and Hybrid SON.

The paper also lists the Use cases and the problems ands solutions for the use cases.

NEC whitepaper on SON is quite recent and it lists the recent standards status:

3GPP has introduced SON items in its standardization path as required features for LTE deployments. Rel. 8 includes the first specifications on requirements, integration with operators’ processes, and identification of main use cases. Rel. 9 is expected to define advanced features, which will introduce self-healing and self-optimization capabilities into LTE. The SON related specifications are driven from the SA5 Working Group (WG) – mainly for architectural aspects– and the RAN3 WG – especially for the optimization of radio functions. Also, Rel. 8 defined the grounding documents for SON: “SON Concepts and Requirements” in TS 32.500, and two main use cases– “Self-Establishment of eNodeB” and “Automatic Neighbor Relation” – in TS 32.501, 32.502 and 32.511.


Zia said...

Even if properly implemented, it would not kill off drive-testing. A network with SON capabilities is certainly better off than one without. Information obtained from performing short & long call, benchmarking drive test give information that would not be avaible with network monitoring capabilities. The information that is available from mobile in SONs is also available with in current OSS statistics but you would need to analysis it with drive test statistics to pin point and understand the precise issue or parameter to be resolved. That said SON will certainly assist in the process of optimization but it wont kill drive test they will go hand in hand.

You have been doing some brilliant work on your blog getting all relevant stories from mobile world together.


Anonymous said...

Could this technology be used to provide greater resiliency in a network in case a site fails? In essence, this is the reverse of the use case described: instead of dropping a new cell into the network (and reducing power levels of neighboring cells) the outage would effectively remove a cell from the network (and increase power levels of neighboring cells). Would this work?

Zahid Ghadialy said...

Yes, the dropping of a cell was just an example, it should work the other way round. If I am not wrong than what you are referring to is known as the 'self-healing' property of the SON.