Voice over Wi-Fi must include Wi-Fi-to-cellular roaming to succeed, according to speakers at the VON Spring 2005 conference.

"People don't want to carry two phones," says Michael Stanford, principal engineer in Intel's Digital Enterprise Group. "And the voice-over-Wi-Fi market is relatively small." So, if given the choice, end users will just carry a cell phone and forgo using Wi-Fi, he says (in any case, with the mass market and subsidies to handsets, cell has a big capital cost advantage).

Moving from Wi-Fi to cell is the key
The key to this type of roaming is how smoothly cellular networks and Wi-Fi networks can hand off calls to each other, with the handoff in the direction of Wi-Fi-to-cellular, he says. This is because it is much easier to walk out of a Wi-Fi coverage area than it is to move out of a cellular coverage area, making it difficult for the phone to know when to set up a cell call to which it can bridge a Wi-Fi call, Stanford says.

Handset vendors will likely push the technology in order to stay competitive with each other, and carriers will likely cooperate because they expect customers to demand the convenience of roaming, he says.

Operators see savings
Service providers are also likely to cooperate because it is less expensive for them to support handoffs between cellular and Wi-Fi than to build more capacity in their cellular networks. So if wireless calls can be handed off to Wi-Fi networks at convention centres or other public areas with large crowds, the Wi-Fi networks can pass the calls off to landline networks for transport. The alternative would be to build more expensive cellular stations at these sites, says David Hagan, president of VoIP software vendor Boingo Wireless.

Of course, this sets up the question of how the operators actually support voice on Wi-Fi, with the operator-centric UMA contrasting with the use of SIP. BT's Bluephone, for example, will roam to Bluetooth for cost reasons.

Voice over Wi-Fi faces other challenges, including adding features to Wi-Fi phones that are already standard with cellular phones, says Joe Rinde, CEO of Rinde Associates. Customers want to have phone books, visual and audio cues for voice mail, distinctive ringing and other common cellular features, he says. Hybrid devices are appearing, but these lacks are among the reasons to be wary of voice on Wi-Fi.

Quality counts
Voice quality is also an issue, says Stanford, but that is not critical. Rinde agrees, saying that customers like mobility enough that they are willing to put up with what is pretty bad voice quality in some cases.

That will be remedied over time with better coder-decoders that pick up on a broader range of speech pitches, making it easier to distinguish between similar sounds that use high frequencies, such as effs and esses, Stanford says.

Voice on Wi-Fi can add apps
The biggest potential advantage voice-over-Wi-Fi phones have over traditional phones is that they can be programmed with new features without involving the maker of the phone, says Rinde. "You don't have to go back and beg Avaya and Nortel to put them in," he says. "What are the killer apps? Nobody knows, but VoIP makes it easier to try."

Vendors need to work out how much power Wi-Fi chews up so people can use phones longer without recharging batteries. "Wi-Fi is a battery hog," says Stanford. This isn't as much of a problem in corporate networks where users are likely to cradle the phones in rechargers when they're not talking.