Last week in Frankfurt, I participated in the second industry conference organized by NGMN. The session opened with speeches by the Hamid Akhavan, CEO of T-Mobile, Vivek Badrinath, CTO of France Telecom/Orange, and Steve Pusey, CTO of Vodafone. Each spoke about the challenges ahead, and some common themes emerged, including royalties, backhaul, and self-organizing networks.Royalties for intellectual property are a key concern. The next-generation mobile network-the Mobile Internet-must support devices beyond simple mobile phones. New devices like the iPhone, of course, and laptops, but the operators also hope to attract innovation from the consumer electronics industry. Imagine a wide range of mobile entertainment devices, for music, video, and games, handheld or in vehicles. Yet the business case for developing such devices would be overwhelmed by the typical royalties of the mobile industry. Current 3G device manufacturers pay over 25% of the retail price in royalties (unless they have patents of their own to balance the negotiation). In contrast, the patent royalties for an MP3 player run only 75 ¢ per unit. Consumer electronics firms are unlikely to introduce mobile products if the business case includes handing over a quarter of the revenue for just the mobile interface, a relatively small component.The NGMN IPR Plenary is seeking to address this concern by asking patent holders to submit anonymously the royalty rates they expect. A trusted third party then aggregates all the royalty expectations and reports back the total. So far, the total still resembles higher mobile rather than lower consumer electronics royalties. And more royalties for older cellular technology (GSM, CDMA, 3G) might also be required if those are included for backwards compatibility. The negotiations continue though opaque layers of technical and legal jargon. Leaving existing royalty models won’t be easy for net collectors like Ericsson and Qualcomm, but such a change would unleash new business by accelerating the Mobile Internet.Backhaul from cell sites is another concern for operators. Historically, highly compressed mobile voice calls required limited bandwidth, and the air interface was the bit-rate bottleneck. Thus, a typical cell site could be connected with just one or two T1 or E1 lines over copper or microwave. Increasingly, laptops cards are multiplying traffic exponentially (and VPN encrypted traffic cannot be compressed), while next-generation radios are removing the air interface bottleneck. Just adding more T1/E1 lines would cause operator costs to outpace revenue. The eventual solution to this challenge seems to require fiber to the cell site, architecturally aligned with Carrier Ethernet. The migration towards this architecture will involve cell-site routers that multiplex circuit and ATM traffic from legacy base stations with new native IP traffic. This change is part of the broader trend toward flat IP architectures for mobile networks.Self-organizing networks promise to lessen operational concerns. Existing cellular networks require much manual configuration and optimization to ensure neighboring cell sites work properly and hand off connections successfully. As next-generation mobile networks grow, such manual tasks would greatly burden the operator. Instead, the radio and network components should interact among themselves, to configure and tune the mobile system automatically in real time. (For historical precedent, consider the migration from IBM SNA networks, which required substantial sysgen configuration, to TCP/IP, which substituted automatic routing protocols.) Self-organizing networks are often associated with mobile ad-hoc networks, but the discussion at this conference flowed more from femtocell developments. To succeed in the mass market, femtocells will require plug-and-play installation by consumers in their homes. Radio and network parameters must be configured and tuned automatically. Presumably, the same techniques can be scaled to macro base stations in the next-generation mobile network.My presentation, Service Delivery in Next Generation Mobile Networks, discussed how mobile operators can introduce new services. It contrasted mobile industry systems like IMS with Internet ecosystems for Web 2.0, drawing on themes covered in previous posts to this blog. (About IMS, Mr. Pusey cited the old joke about mixed feelings, “like having your mother-in-law drive off a cliff in your brand new Ferrari.”) I described new capabilities for enterprise customers, like VPN and web conferencing, and suggested consumers will extend social networking using mobile. Anthony Berkeley, Alcatel-Lucent LTE Business Strategy manager, offered the wry observation that everyone at the conference was too old to anticipate the popular services on the next-generation mobile network, as teenagers will most likely set the trends. With that, I was ready to return home to ask my children about the future of the Mobile Internet.