SON: Where is it heading?

NEW YORK, USA: There are several exciting and potentially game-changing innovations underway for communication networks and the software architectures that support them. The self-organizing network (SON) is only the latter. It is neither as new nor exciting as Software Defined Networking (SDN) or Network Function Virtualization (NFV).

That SON should command less attention than SDN or NFV is unfortunate; but also may be a positive sign that CSPs are now truly focusing on developments that directly benefit the customer experience rather than the bottom line, which SON can certainly help them deliver. Some would argue that SON also improves the customer experience by ensuring that capacity meets demand-but Stratecast views SON primarily as a streamlining and optimization exercise, for now.

Also, SON will not get the inter-departmental push the other technologies get because it is not something CSPs can easily charge for. It is not a revenue generator. Besides, CSPs are in no hurry to hand over the planning, configuration and maintenance of their networks to a computer. Still, SON will come. It is a significant step in a long-desired direction. It will not come all at once, but will proceed tentatively from the largely manual process that exists today, to a semi-automated process- where Stratecast believes it will linger until some brave or desperate CSP goes fully automatic.

The underlying concepts of SON are not new; programmability and automation have been the goal of network engineers for decades. But the tools necessary for achieving them have only recently evolved to the degree at which they become plausible. As described in this report, the terminology is evolving with it.

Self-Organizing Network guidelines were first introduced as part of the 3GPP's LTE specification, in 2008, for the radio access portion of the network known as the RAN. Although the term Self-Organizing Networks remains the official terminology within the standard, most operators and vendors that Stratecast spoke with for this report refer to SON as "Self-Optimizing Networks."

According to the 3GPP standards, optimization is just one of three functional subgroups of Self-Organizing Networks. The subgroups of SON are self-configuration, self-optimization, and self-healing. Ultimately, all three subgroups optimize the network in their own way. They are intrinsically connected.

- Self-Configuration - The self-configuration function, or subgroup, focuses on the pre-service deployment of various network elements, primarily in the radio access network (RAN). The goal is to allow base stations and other elements to establish connectivity, and configure parameters on their own, as they are switched on. In this manner, they discover their place in the network and the configurations of their neighbor and adjacent cells; which, in turn, automatically adjust to their new neighbors. While this initially does not eliminate the need for engineering to design and plan the overall network configuration, it should significantly reduce the time it takes to prepare, install, turn up, and configure new elements in the network This will become increasingly important as small cells and heterogeneous networks proliferate, as is expected.

- Self-Optimization - The self-optimization function, or subgroup, is designed to improve coverage by maximizing the utilization of network elements, and conserving the power to run them. It concentrates on the growing number of parameters that need to be set and reset in response to changing requirements in traffic, time-of-day demand, emergency coverage, environmental conditions, and the health of other network elements. Eventually, it will include the automatic reconfiguration of antenna-tilt settings for a new generation of radio.

- Self-Healing - The self-healing function, or subgroup, will respond to service degradation or system failures by automatically adjusting parameters, configurations, or physical tilt in neighboring nodes to compensate for lost coverage. This self-healing function could change the dynamic of troubleshooting networks, and improve mean-time-to-repair and service quality. It could also reduce the severity and duration of service affecting outages.

As promising as SON is-within each of its components-network operations departments are not ready to manage a fully autonomous network. Network planners are not ready for self-configuration. Network engineers are not ready for self-optimizing networks. As one North American operator said, "We don't know how to run that kind of network."

Most of the concern over rushing into SON stems from a reluctance to embrace real-time modifications to active network configurations. Of primary concern is losing the point-of-reference for troubleshooting when something goes wrong or a series of parameter changes does not have the desired effect.

CSPs are wondering if they might have to completely rethink the network maintenance processes that have stood the test of time for more than fifty years. Stratecast believes that before CSPs get comfortable with network-wide, real-time network configuration changes, they will apply SON to localized portions of their networks, or through "distributed SON"-one of three SON architectures.