Self-Organizing Networks: Making a Good Concept Real

Global mobile data traffic tripled for the third year in a row in 2010, and, according to Cisco, the total mobile data traffic will grow to 6.3 exabytes per month by 2015, a massive 26-fold increase over 2010. This unprecedented demand for bandwidth puts huge strains on all parts of the mobile ecosystem, and to address it a plethora of technologies are coming on line, such as EV DO, HSPA +, LTE, and now LTE-Advanced. Additional benefits come from multi mode devices, intelligent small cells (which include distributed processing and smart antennas, such as femtocells), and heterogeneous network topologies that spread the load across various types of base stations. While helping address bandwidth issues, these innovations come with a downside: highly complex networks that are expensive to manage.

Such complexity and expense were taken into consideration during the formulation of LTE. This led to the development of Self-Organizing Networks, or SONs. The goal of a SON is to automate network operations and management in order to minimize operational costs.

Managing SON Complexity

Good planning, configuration, integration, and management of network elements are imperative for efficient performance. But they require specialized expertise, and managing them manually tends to be expensive and error-prone. Automation through SON makes the processes more efficient, and reduces costs and errors.

The initial SON innovations focused on self-configuration of networks through means such as plug-and-play hardware, self-configuring radio devices, and automatic neighbor lists. These had big benefits for network planning and network deployment, helping simplify the amount of pre-planning required, and also reducing the overall deployment times. Network improvements allowed further OPEX reductions through automated features such as neighbor optimization, handoff optimization, QoS optimization, and interference reduction. Also, self-operating features such as automatic inventory, cell outage detection and compensation, and multi-vendor trace have helped reduce the level of monitoring and adjustments required.


Evolution of SONs

Over the past three years as LTE has developed, the concept of SONs has matured significantly also. The goal now is to extend the impact of SONs beyond simple expense reduction, and to address things such as call quality and network optimization. These goals have gained steam in the latest 3GPP releases, and a few of the relevant enhancements and their implications are discussed below.

  • Mobility Load Balancing: Load balancing between 2G, 3G and 4G networks to optimize the use of air interface capacities of pooled radio frequency resources. This helps in distributing load and optimizing the number of base stations.
  • Mobility Robustness Optimization: This facilitates detection and resolution of radio link failures by providing information for possible corrective measures, and helps in improving QoS.
  • Inter-cell Interference Coordination: The enhanced interaction between base stations helps identify and mitigate interference by improved coordination in the access network.
  • Coverage and Capacity Optimization: This involves evaluating network performance using key performance indicators from base stations, as opposed to expensive drive tests.
  • Energy Saving: Various energy-saving mechanisms are being considered, such as autonomous switch-off in base stations. These help reduce the energy footprint of the access network.
  • Self-Healing: Identification and isolation of faults need to automatically trigger appropriate recovery actions in order to minimize downtimes and the need for manual intervention.

Take a Broad View of Your Network

Theoretically, SON implementation could be considered as a set of isolated use cases, each of which presents a distinct benefit. However, in the real world, implementation of SON involves a connected scenario where many use cases will interact and influence each other. This means several complex interaction scenarios must be taken into consideration at once.

An additional complication is that the processing and response times influencing - and being influenced by - a SON vary drastically. For example, a radio response management may respond in milliseconds, while an OSS-based optimization may take hours. Therefore, SON implementation must take a broad view of network operation in order to see the desired improvements. Moreover, the SON solution should be easily scalable and upgradable to accommodate future innovations.

Further, SONs target a whole range of parameters such as antenna tilt and MIMO configuration, and choosing key parameters that address multiple scenarios is important to reduce complexity. SON can be implemented using either centralized, distributed, or hybrid architecture, and each has its own pros and cons. The best way is to strike a balance is by dividing functions on the basis of complexity, response time requirements, and parameters for optimization.

Making a SON Work for You

Besides the technical considerations listed above, the cost of SON implementation itself needs to be optimized relative to the amount of OPEX savings. This requires that the SON is compatible with existing operations and maintenance workflows, since creating a whole new set of workflows is often costly and might defeat the point of the SON.

In short, implementation of SON for LTE is a highly complex affair. While SON functionality has been specified in the 3GPP Releases, the standards merely ensure interoperability and there is very little guidance on actual implementation. There is still significant room for differentiation between vendors in the area of optimization algorithms and SON accuracy/efficiency, and it is the choices made by equipment vendors in the development of these features that affects the actual benefits seen by the carriers.

Leave a Reply

Your email address will not be published. Required fields are marked *