A commonly accepted definition of capacity is the one provided by Shannon which states that capacity is the maximum achievable set of rates in multiple access channels with an arbitrarily small probability of error. As this metric represents a bound in performance, in practice, the sum of the transmitted data rates (downlink) or aggregated data rate is used.
However, with the increased availability of new services in wireless networks, user perceived quality or QoS is now also included in many capacity measures. For instance, voice services have long been designed with a probability of error (non connection) ranging from 1% to 3% In the data centric world, the system capacity could be defined as the maximum aggregated data rate subject to the constraint that the average experienced quality of all flows in the system should be fulfilled according to a given target.
As “average” experienced quality we can mention the “average” delay of all transmitted packets or the “average” packet throughput. Since the required “average” experience varies across different services, the traffic mix chosen by the Operators will have a strong influence on the final maximum aggregate data rate that will be required and smart phone will further complicated the situation with their new user behavior pattern.
The aim of LTE capacity dimensioning is to obtain the PS throughput supported in the network based on the bandwidth available and channel condition of each user. A high level summary for capacity planning process and input requirement is listed in the diagram.
Examples of “Scenario Parameters” and “Equipment Parameters” are also seen in figure.
Most of these parameters are similar to those used for 2G/3G network dimensioning and by carefully considering the contribution of all these parameters, network planning engineers can determine which customer service level can be met.
Nevertheless, the arrival of smart phone, which has completely different behavior compared to feature phones, is going to add a new level of challenges to planning engineers. They frequently changes state between “idle” and “connected”, its fast dormancy feature forces the terminal to switch to an “idle” state every six to eight seconds in order to save battery power, and the service heartbeat mechanism periodically communicates with the application server. According to signaling statistics of operator, one smart phone creates 14 times the signaling load of a feature phone.
In addition, the increasing popular level of applications like twitter will hasten the evolution of customer behavior and traffic model in the next few years. Average subscriber usage at busy hour has rapidly increased from the low 10kbps (since R99/1xRTT) to be in the mid to high 30kbps right now.