Wi-Fi/Cellular Handover - The State of the Market

In the mid/late 2000s, one corner of the mobile universe was focused on (some might say obsessed with) the seamless, in-call handover over a voice session to (or from) Wi-Fi  from (or to) the cellular network. 

At the time, this was known as “fixed-mobileconvergence”.  There were companies trying to solve this from the network angle (Bridgeport, Outsmart, NewStep, Convergin,…), others from the enterprise side (DiVitas, Tango, Agito,…).  There were even industry associations like the FMCA and MobileIGNITE.  Nearly all of these companies/approaches are dead and gone. 

In the end, there was and continues to be, only one technology which supports seamless handover, 3GPP GAN (aka UMA) technology. 

Originally T-Mobile marketed this as “HotSpot @Home”, withsome very funny commercials.  Orange, Rogers, BT and others all has similar services and seamless handover was a key requirement for suppliers to this service.  Watch a PhoneScoop video of the cool Nokia 6086 phone from T-Mobile showing in-call handover from 2007. 

Then in 2010, T-Mobile re-launched their GAN-based service under the brand name Wi-Fi Calling geared towards smartphones – with one technical difference, Android devices with Wi-Fi Calling did not support handover. 

Outrage ensued!

Well, not really.  In fact there was almost no public comment (e.g. Twitter complaints) from T-Mobile’s user base.  While thrilled with the benefits of Wi-FiCalling (specifically coverage), users were ambivalent to handover. 

Let me offer some insight as to why this may be the case:

• Dramatically larger Wi-Fi coverage area.  Ironically when a phone needs to worry about handover, any dip in Wi-Fi signal immediately becomes cause for concern.  The signal strength threshold for deciding to hand over needs to be sufficiently high so the call doesn’t drop, resulting in a rather modest range (or radius) from the Wi-Fi AP.  People found their phones handing out to cellular in corners of the house/office – sometimes in places where there wasn’t sufficient cellular coverage to ‘catch’ the hand-off.  Whoops.  By eliminating handover, the phone is forced to focus on maintaining a connection to the Wi-Fi signal, thus dramatically increasing the effective coverage area. 
• Differentiated billing.  By inserting a ‘seam’ between Wi-Fi and cellular, T-Mobile was able to bill differently for calls over Wi-Fi.  In fact, with the new service, domestic calls over Wi-Fi aren’t billed at all.  This enables T-Mobile to be competitive in locations (like the home) where there is a lot of telecom competition – from fixed lines to Skype and other Wi-Fi-based calling services – and maintain a premium for outdoor/cellular service.  (A billing benefit also provides incentive for users to turn on and connect to Wi-Fi, which drives more data offload, but that’s a different story…)
• More devices supported at a lower cost.  It stands to reason that handover is more complicated than no handover.  The millisecond timing required for handover requires low level of integration.  GAN handover is implemented at the silicon layer, GAN without handover is implemented in the OS layer.  The result is much faster time to market for a broader range of devices ultimately at a lower cost.  

To sum up the state of Wi-Fi/cellular handover today?  3GPP GAN supports it in the specification, it’s been proven in the field, but it seems that from a commercial perspective, it’s time has passed.