Why Apple Must Enable FaceTime On LTE

Ever since Bell Labs unveiled its Picturephone 50 years ago, people have been captivated by the idea that calls would migrate from voice only to voice and pictures simply by merging the two most significant pieces of consumer technology of the day: telephone and television. Achieving that vision has been fraught with frustration but, finally, we're oh so close. 



Yes, webcam-equipped PCs have been able to augment chats with video for a while. But it's taken the smartphone to really deliver on the video-calling promise. With all due respect to Skype, it's FaceTime that has brought video calling to the masses--at least for the millions of people using iPhones and iPads. By integrating FaceTime into the iPhone phone OS (oddly, it's still a discrete app on the iPad, Mac, and iPod Touch), Apple has turned video calling into just another communication channel, an alternative to a voice call or text message.

But it's still a tease. While FaceTime has succeeded in lowering the technical hurdles to video calling, it hasn't eliminated them. It's neither ubiquitous (you can't use it on a cellular network, just Wi-Fi) nor federated (you can't call someone without an Apple ID on a non-Apple device). Fortunately, both of these roadblocks are surmountable--if Apple is serious about fulfilling Steve Jobs' promise.



First, the network. Apple has limited FaceTime to use on Wi-Fi for fear that that bidirectional video streaming would overwhelm 3G networks and destroy the user experience. Touche. Who wants a video call that breaks up and rebuffers like DirecTV in a thunderstorm? However, improvements in carrier networks and codec efficiency have greatly mitigated these concerns. While Apple claims that HD video requires a minimum download bandwidth of 1 Mbps, FaceTime calls from mobile devices, which only have a VGA front-facing camera, would use far less. An Aruba white paper, "Apple FaceTime On Multimedia-Grade Aruba WLAN" [PDF], goes through the bandwidth calculations in detail and concludes: "In planning for widespread FaceTime deployment, a representative bandwidth requirement would be around 350 Kbps in each direction, for a total of 700 Kbps per call, including both video and audio streams." And that's being conservative. Aruba's actual measurements show an average unidirectional data rate closer to 300 Kbps. Another test on a jailbroken iPhone using FaceTime over 3G found upload and download rates running about 3 MB per minute, or about 410 Kbps. 



Such bandwidth is within the capacity of current wireless networks. SpeedTest on a Verizon iPhone 4S and third-generation iPad shows calls on Verizon's legacy 3G network clocking in at 400 Kbps to 500 Kbps down and 250 Kbps to 400 Kbps up, which on the low end just about matches Aruba's estimates. For a review of the new iPad, I tested typical speeds on AT&T's HSPA+ network (LTE isn't yet available in my area) at 3 Mbps to 5 Mbps down and 1.5 Mbps to 3 Mbps up--many times what a FaceTime call would require. And, of course, LTE blows these numbers away, as discussed in a recent report. LTE speed tests on the new iPad are running at 15 Mbps to 30 Mbps in each direction, far faster than many home broadband links.

Yet Apple still refuses to allow FaceTime over anything but Wi-Fi links. 


When people are already blowing through gigabytes of data streaming March Madness games, it's time to end the madness of blocking person-to-person video that would chew up less than 100 MB over a half-hour. But there's call for optimism since developers hacking the latest iOS 5.1 code have found signs Apple is planning to enable FaceTime on LTE. Here's hoping the company relaxes the policy further to allow it on any network capable of 300 Kbps in each direction.



Federation is a thornier problem. Video calling should be like voice and email, where you needn't worry about what carrier or service your counterpart happens to use. And in fact, on announcing FaceTime almost two years ago, Steve Jobs claimed Apple would "take it all the way. We're going to the standards bodies, starting tomorrow, and we're going to make FaceTime an open industry standard." But there's no sign of that happening. Although FaceTime itself uses a raft of standards, the amalgamation of these into an open, cross-platform application requires an open integration effort, much like what Mozilla and WebKit did for browsers. Unfortunately, there's no sign of FaceTime working with Skype, or any of the other walled-off IP telephony fiefdoms, anytime soon. 



The incompatibility of various video-enabled communication channels, whether FaceTime, Google Talk, or even Facebook, is driven more by business strategy than technical merit. It doesn't stem from some arcane incompatibility in data formats, but rather an inability of disparate systems to interoperate: to authenticate and find users, and route traffic across networks. Email solved this problem decades ago. It's not too much to expect videophone systems to solve it now. Apple should take the lead on two fronts. First, eliminate the restriction of using FaceTime on cellular networks. There's no reason the iPhone can't be today's wireless Picturephone. Second, work with standards bodies to develop and proliferate a means of initiating and routing real-time communications--not just video, but text and packetized voice (i.e., VoIP)--across different service providers. We need a 21st-century version of the telephone number that works across the Internet. 



As compelling as FaceTime is, it only highlights the obstacles we've yet to overcome on the road to convenient, ubiquitous, mobile video communications.