Voice over WiFi has been steadily moving forward as a way for wireless operators to offload traffic; for OTT providers, like Skype and FaceTime, to expand their offerings; and for cable operators to launch a mobile product.
In order for the cable contingent to compete as an equal option, however, the service has to be of equal or better quality, says Sean Yarborough, senior director for product management, strategy and marketing, Spirent Communications. He offers five best practices for testing and measuring a voice over WiFi service to ensure a good result.
1. Measure what matters most to the consumer. To do this, an operator must measure the call performance as experienced by the subscriber. Is there an echo? Distortion? Low audio level? But how many calls actually connect is also significant. "If I made 100 calls and only 90 connected, I would be concerned and would probably switch providers," Yarborough said.
Related factors are how long it takes a call to connect. Two to three seconds is a good target range, but sometimes voice over WiFi experiences longer delays, maybe eight or nine seconds, due to handovers between hotspots. In these cases, the consumer often hangs up and assumes the call never went through.
Consumers also may be experiencing dropped calls. "If they stay on the call for only two or three minutes before it is dropped, they might switch providers. Measuring what matters from the end user perspective is the first thing to focus on," Yarborough said.
2. Measure the service end-to-end. Focusing on the health of the network is not thorough enough. The operator must address performance from the consumer's device to another consumer end point. "Jitter, latency and dropped packets could be low, but calls could still not be connecting," Yarborough said. "You need to test at the end point of the network to understand how the consumer is experiencing the service."
3. Compare the service to the competition. Sometimes a provider might think they are ready to launch, but when matched against a 3G service or an OTT provider, the performance may still be lacking. Benchmarking should be done by actually testing the services in a real-world situation, like a coffee shop. The devices should be the same to take that variable out of the equation. Voice over WiFi calls can be compared to 3G or LTE calls for example, and measured using POLQA, a next-gen mobile voice quality standard.
4. Load also can impact service. Voice over WiFi could be running great in downtown Washington DC, for example, with 100,000 consumers. But if calls jump to 1 million per day and users are not doing just voice, but also streaming video, will the service still perform well? "The ability to test under different types of loads is important to make sure the network is ready," Yarborough said.
5. Monitor 24/7. Once the service is up and running, and the load changes and number of subscribers increases, there must be a way of actively alerting the operator to problems. "They may start with one implementation and six months later deploy a new access point," Yarborough said. "What is the effect of this change on those services? They need to continuously look at performance on a 24/7 basis."