The VocalTec Blog

iPhone 4 and FaceTime – what does it mean?

As expected, the new iPhone has a front-facing camera for video calling. But unlike many similarly equipped 3G handsets it does not support video calling functionality on cellular networks. This positions FaceTime more as a video chat application, similar to Skype. In fact, the underlying open protocols mentioned by Jobs (referring to the acronyms on the accompanying slide as “alphabet soup”) are used in many PC-based video chat clients, including Apple’s own iChat AV.

Cellular video calls

Video calls in 3G cellular networks today use the 3G-H.324M standard. Basically it is an evolution of ITU’s H.324 standard (which in turn builds on H.320 and H.323), multiplexing audio, video and control protocols into the same TDM circuits normally used for voice calls.

Note that H.324M does not use IP transport: it is circuit switched technology. Because of this it does not have to deal with NAT traversal or the quality of service issues that IP-based real-time communication can suffer from. Moreover, since subscribers are addressed using the same E.164 phone number users already know and use for voice calls on a daily basis, the barrier for users should be pretty low: all it takes is compatible devices on either end (and a willingness to pay the per-call fees, which may be more than for a voice-only call).

Video chat

Video chat, on the other hand, has its roots on the Internet, and is therefore inherently based on IP protocols. Most video chat client applications today are PC-based, and many of them originally started as Instant Messaging or voice chat applications. There are various protocols, some open and some proprietary, and many audio and video codecs. Apple uses SIP signaling, H.264 video and AAC audio, over RTP or SRTP. The other protocols in the “alphabet soup” deal with NAT traversal; in particular the support of ICE is important, as it should take care of all but the most obscure network topologies.

In order to use a video chat application, users typically have to sign-up with one of the networks. Since not all networks use the same protocols (and even if they do, they sometimes artificially restrict ‘federation’ with other networks), users may even have to sign up with several networks in other to communicate with friends that happen to be in other networks.

Addressing issues

This fragmentation of the (video) chat networks leaves the mobile carriers with a key advantage: all you need to know to call someone is their phone number, regardless of the service provider the called party is using. Attempts to make VoIP communication use the same addressing scheme, in the form of ENUM, have failed to see any major deployments. Services that allow the user to sign-in once and connect to users in various networks through gateways make life a little easier, but it is still necessary to set up a client with a username and password, and to add IM screen names to all contacts.

While Apple’s FaceTime appears to contribute to more fragmentation (for now only iPhone 4 users can call one another), it does manage to hide the addressing complexity of typical chat networks: there is no need to enter account details, no need to add screen names to contacts, and video calling is an integral part of the phone dialer interface.

Open standard?

Jobs announced that Apple will ‘open up’ FaceTime as an open standard. Since much of the underlying communication protocols/codecs are already well established standards, the only ‘missing link’ for now appears to be the ‘zero-configuration’ aspect of the system. It will be interesting to see what Apple will come up with, but if this is something that can be made to work on other handsets too (Android, anyone?), then the future of mobile video chat may be bright, especially once the service works over 3G/LTE data connections as well. However, it is difficult to see what’s in it for Apple’s carrier partners: have they resigned to the idea of being a ‘data pipe’ for all IP-based services, and are they ready to let go of their voice calling stronghold?