The mobile TV market seems to be settling into two camps. AT&T and Verizon Wireless are championing the dedicated network, specifically Qualcomm's MediaFLO, which uses broadcast technology over new spectrum to send high-quality TV to dedicated devices. On the other side is Sprint, which — with its plentiful PCS spectrum and high-capacity 3G network — is opting to stream its TV services over the cellular network.

Each approach has its advantages. MediaFLO and other broadcast technologies, such as digital video broadcast-handheld (DVB-H), provide dozens of channels of high-quality video without using any of the operator's resources. But those services can be received only over purpose-built devices containing a DVB-H or forward link only (FLO) chip. Unicast services over the existing cellular network not only allow any device with a media player to access mobile TV, they also allow for video-on-demand services as well as a level of interactivity only achievable on a two-way network. But unicast networks use the carrier's existing network and spectrum, and there's perhaps no greater bandwidth hog than TV.

Looming on the horizon is an alternative technology that is still at least a year away from commercialization: multimedia broadcast multicast service. MBMS is a compromise between the unicast and multicast models. Like unicast services it uses the carrier's existing infrastructure and spectrum, but unlike dedicated unicast, which sets up an individual point-to-point link for every video user, MBMS has a broadcast mode in which it can place all users watching the same video content on a single channel, using the 3G network as a broadcast network.

While MBMS often is mentioned in discussions of other mobile TV technologies, so far there has been little interest in it in the U.S. MBMS is designed to work with UMTS networks, which are limited to AT&T and T-Mobile domestically, and the commercialization horizon of the technology is still distant, whereas MediaFLO is readily available and DVB-H networks are expected to launch soon. Nokia Siemens Networks officials believe, however, that there is a latent market for MBMS in the U.S. as TV services evolve and the technology is readied for launch in 2008. Clay Simmons, solutions manager for the service core and applications for Nokia Siemens, said that MBMS won't necessarily be a disruptive technology to the dedicated multicast networks, but rather a complementary technology that will allow wireless carriers to offer TV to a larger subset of their customer bases.

“I don't think MBMS will be a regionally specific technology,” Simmons said. “In fact, the North American market may very well be the major target market for MBMS.”

MBMS works by creating a tiered-delivery video-distribution system over the UMTS network. A typical unicast network like Sprint's treats video packets like any other data packets, establishing a direct and dedicated link between the content server and every end device. But not all video links are created equal, Simmons said. The MBMS network aggregates all video streams in a media server sitting behind the Gateway GPRS Support Node, meaning each channel of video is received once by the base station no matter how many users on the network are viewing that channel. But, in contrast to a broadcast network, the media server relays only the channels that are actively being watched on the network. The end result is that the network's overall transport capacity and radio capacity are conserved, Simmons said, as unviewed channels don't traverse and those channels being watched by multiple subscribers are sent as a single transmission, not as a multiple threads.

But MBMS doesn't just optimize the network for unicast TV, Simmons said. The architecture alternates between unicast and broadcast modes depending on network conditions and the number of users on the network. If only two subscribers are tuning into TV broadcasts in a particular cell, the base station can simply allocate separate unicast channels for those two users. But as more TV subscribers move into the cell demanding more resources, the network can shift them all onto one high-power broadcast channel. Capacity can be doled out or reigned in depending on dynamic demands for TV, data and voice on the network.

MBMS's ability to handle the ups and downs of unicast and multicast networks makes it a step toward more interactive TV networks, said DeWayne Nelon, president and CEO of Ortiva Wireless, which has developed a technology that changes a unicast TV transmission's encoding as network conditions change. Nelon said multicast networks only serve a portion of the potential TV market. Services such as video-on-demand and interactive TV require unicast networks for reliable delivery. While multicast may supply a core selection of video content, a unicast element will be necessary to provide high-value, revenue-generating TV applications, Nelon said.

“Services like MediaFLO require you to watch what Qualcomm wants you to watch when they want you to watch it,” he said. “Multicast is trading scarcity of capacity for scarcity of content.”

That may be the case, but persuading any U.S. carrier to build out an MBMS network or a new TV network of its own will be difficult, said Roger Entner, wireless analyst with IAGR. With MediaFLO already available, Modeo's DVB-H network in trials and HiWire's upcoming launch, there will be several choices for operators, none of which involve any of their own capital expenditures. Entner pointed out that 80% to 90% of the AT&T and Verizon Wireless channel lineups will be the same. Multicast gives little opportunity to differentiate, he said, and while unicast networks may give service providers more leeway to offer higher-value service, it will still be hard to justify the capital investment and drain on capacity.

“They're looking for the most efficient way to use their spectrum, and television is the biggest bandwidth hog of them all,” Entner said. “Anybody who expects mobile TV to be a determining factor in choosing an operator will be disappointed.”