4 reasons to use real data applications to validate user experience
We all know that data traffic volumes are increasing exponentially, while at the same time user expectations of quality of experience are higher than ever. With network operators now frequently bundling in subscriptions to audio- or video-streaming services with their premium data packages, it is becoming ever more difficult to predict the volume and type of traffic that users will generate, particularly as the data used by the bundled services may not actually be part of the monthly data plan. Just because this data is ‘free’ doesn’t mean that the user will tolerate a poor service, so the way in which network experience is validated is growing in impartance.
Here are the four key reasons why ‘Real Data Applications’ emulation is the best way to ensure that subscribers receive the user experience they expect:
1. Quality of Experience (QoE) means more than just network integrity
Although the two terms are sometimes used interchangeably, QoE does not mean the same as Quality of Service (QoS). As shown schematically in Figure 1, QoS includes network performance, terminal performance, and applications factors, while QoE adds on top of this subjective human components such as customer care, billing, and user perception/emotions. In other words, QoE is measuring the performance of the application and QoS is a performance metric of the network.
2. Applications are not all equal in their behaviour, nor in the demands that they place on the network
Both the size of the packets and the frequency with which they are sent, as well as whether the traffic is asymmetric or almost symmetrical, will affect their impact on the network. For example some social networking applications such as Twitter and Skype are characterized by the transmission of small data packets in both uplink and downlink. This can cause the handset to switch frequently between idle and a connected state, and requires more extensive signalling during handover. If it is kept in the connected mode for longer, then this impacts power consumption. On the other hand, video streaming and web browsing require the transmission of larger packets, with a consequent demand on downlink data capacity. Table 1 summarises the key characteristics of the main types of applications traffic.
3. The achievable data rate from a base station (eNB) is inversely proportional to the number of users connected to it at a particular time
Because the data rate is not a fixed quantity, it is essential to test the scalability of the network by loading it with large numbers of users working with a realistic mix of applications.
4. There are different ways of hosting applications data
Some applications data such as video may be hosted or cached by the operator on its Content Delivery Network (CDN) to improve download speeds, while the rest will be carried externally. This also needs to be emulated in order to establish meaningful QoE metrics.
Today’s increasingly intelligent wireless networks police and prioritise different types of traffic in a different way. In order to test the capacity, validate new functionality, and to optimize network performance, it is crucial to evaluate Quality of Experience – and its components – at both an individual application level and an individual user level, if the test scenarios and results are to be meaningful in terms of how the network will perform in service, and how satisfied its customers will be.
Click on the button below to find out more about our Validation solutions