There are about 4 billion mobile devices in the world today, said Len Lauer, of which about 830 million are 3G today, but the vast majority of the remainder will be migrating to 3G in the next five to seven years. The COO of Qualcomm was speaking with Stacey Higgenbotham a staff writer with GigaOM during the Mobilize 09 conference held in San Francisco on Thursday September 10th defending the position that there’s plenty of life left in 3G now that 4G has begun to appear in the market. I was a fly on the wall listening to their conversation over the Internet thanks to LiveStream.com.
As with previous wireless technology transitions, 4G handsets will be a multimode devices switching from 4G to 3G when outside a 4G service area, which for the near future will be limited to dense metropolitan areas such as LA, New York, London, Paris, Mumbai… Lauer referred to the CTO of Verizon Wireless, Tony Malone, speculating that 3G networks would be carrying voice for the next decade. The major difference between wireless generations is in the bandwidth provided users: from 56 kbit/s up to 114 kbit/s for the previous generation 2G networks, up to 348 kbit/s to 7 Mbit/s for 3G depending on whether the handset was mobile or stationary, respectively. Next generation 4G will see bandwidths of 100 Mbit/s for mobile operation and 1 Gbit/s for stationary. These 4G data rates will certainly be required if social networking users begin capturing and sharing HD video and still images on their favorite on-line community.
However, bandwidth expansion is not a well-ordered process, more a series of fits and starts as the troubles plaguing AT&T in the wake of the new iPhone 3GS launch attests: dropped calls, extremely slow connections and a generally unpleasant experience. This contradicts the incredible, amazing, awesome… experience touted at the Apple rollout. Today the problem exists in the backhaul, the link between the cell tower and the wireless carrier’s network core, which routes the call to its final destination. AT&T needs to invest in boosting its backhaul capacity and it’s investing elsewhere—increasing the number of cell sites, etc.—to the chagrin of Apple.
And you can kind of understand AT&T’s behavior. The Smart Phone represents only 15 percent of the total market and contrary to the popular notion, a good many iPhones are being brought by the well-heeled consumer that can afford the acquisition cost as well as the $100/month subscription. According to the Wireless Association CTIA, there were 270 million wireless subscribers in the U.S. at the end of last year and 85 percent of this total, 229 million subscribers, are also demanding service providers’ attention.
What’s the problem with wireless carrier backhaul? As bandwidth speeds on cellular networks increase and packet-based mobile services attract more users, traffic growth is making traditional backhaul network designs unmanageable. To accommodate 3G build-outs, service providers are reengineering the mobile backhaul infrastructure using Carrier Ethernet technologies. The problem with Ethernet has been that packet delivery was non-deterministic. If the network got overloaded packets could be dropped, something voice packet cannot tolerate.
The problem is now being addressed with Ethernet IEEE 1588 Precision Time Protocol (PTP) and Synchronous Ethernet ITU G.8261, two standards that make Ethernet packet delivery more deterministic. Used together, the two achieve a high level of frequency synchronization a common defined time. Realistic studies in large-scale deployment scenarios indicate an accuracy of 50 ns—more than sufficient to ensure voice packets are delivered as effectively as the existing circuit switch delivery used today.
Still the technology may be available, but the will to implement lies with the service provider. Their concern as Lauer points out is getting a sufficient return to justify the investment. Lauer observes that for the service providers to reach a larger subscriber base, they have to reduce monthly access costs, while making larger infrastructure investments. It’ll be interesting to see how they solve the problem.
No comments:
Post a Comment