I was at the LTE voice summit in London last week and was confronted with what for me is the voice-QoS paradox.

The repeated message is that VoLTE beats any VoIP service because it provides QoS . And it’s true. But if the message to the user is your mobile network provides you voice with quality, then why deploy VoWiFi?

The challenge is not only that QoS mechanisms over Wi-Fi are mostly not available, not predictable or unreliable; but that VoWiFi is being provided via “untrusted access”, that is, access from a third party provider over which the operator has no control. This is very controversial because in case of a bad experience, like a dropped call, who will the user blame? The operator or the Wi-Fi provider? Both?

So with this situation, why deploy VoWiFi at all? VoWiFi requires new elements in the core network (the so called ePDG, a security gateway) and device to network interoperability tests (a pain for operators).

Well there are also two good reasons for deploying: indoor coverage extension and reducing costs while roaming. Both are known reasons but what goes unobserved is that both are a Wi-Fi last strategy: use Wi-Fi only when nothing else works. Combined, not to forget, with device native support (so no apps), VoWiFi does add value to the user experience, independent of the lack of QoS on the access. (there is interesting data in this whitepaper from Ericsson here)

Operators are taking care that they launch VoWiFi as a Wi-Fi last option and they still promote their main voice service (VoLTE) as a much better experience against VoIP. This is a very significant difference from the 2012 Wi-Fi offload strategies in which data was offloaded to Wi-Fi always when available (with the message that cost per bit was cheaper than LTE)[It is to note though, the technical difference that Wi-Fi offload was mostly intended as “trusted access” unlike VoWiFi]; and the complete opposite what others like Google FI promote, the so called Wi-Fi first operators.

At the conference there was an interesting debate about VoLTE roaming and a pleasant surprise: operators are using RCS to add contextual information to voice services with examples from Korea and Germany. I believe contextual propositions make most sense and it is advisable for operators to position themselves in this area (also with a WebRTC strategy). The key for RCS success in this area (after relative defeat in IM) is again going to be native support in Android devices and that’s where Google acquisition of Jibe this week starts to make sense.

The picture is from Sacha Fernandez with Creative Commons license.