SAN FRANCISCO – With demand for mobile data outpacing forecasts, the industry urgently needs to rally around an effort to find more spectrum. It also needs to set standards for rapidly evolving antenna technologies trying to cover an increasingly broad and fragmented set of frequencies.
Those were two of the messages from the recent conference here of the NGMN Alliance, a group of cellular operators and systems vendors. Today’s budding 4G Long Term Evolution (LTE) networks are using a broader set of frequencies than any previous cellular services to keep up with demand, putting pressure on technologies all across the food chain--and the outlook will only get more challenging, said speakers here.
Mobile data traffic has outstripping projections, catching spectrum planners flatfooted. Earlier this year, the World Radiocommunications Congress (WRC), the group chartered to help plan global spectrum use, said the industry urgently needs to study short- to medium-term needs for cellular spectrum due to the heavier than expected demand.
“The task is to do the study to find out how much spectrum we need and which bands we need for it,” said Michael Kraemer who heads a spectrum task force at NGMN.
The International Telecommunications Union (ITU) created Joint Task Group 4-5-6-7 to do the work and develop a proposal for next WRC meeting to be held in 2015. “The mobile industry needs to be very vocal in this group to make sure the group comes up with the right numbers and bands,” said Kraemer.
Although 2015 seems like a long way off, time is actually short, he said. The ITU group must submit its proposal in late 2014, and that means its decision on what spectrum to request must be made as early as mid-2013, he added.
Some of the studies supporting those decisions are already in progress, including one on use of frequencies below 790 MHz due by the end of this year, he said.
In the meantime, operators are using an increasing diversity of bands to deliver services. Another NGMN working group is preparing on a document recommending use of more than 10 frequency bands for LTE and HSPA+ services by 2015, said Javan Erfanian, a Bell Mobility technologist chairing that group.
Alex Jinsung Choi, head of technology strategy at SK Telecom, gave an example of the pace of mobile data growth. The Korean operator has 3.1 million LTE subscribers now and expects that to double by the end of the year and continue ramping rapidly through at least 2017.
SK’s average LTE user consumes 1.4 Gbytes of data a month, a figure that is also rising. “So we have to prepare for capacity increases,” he said.
Like many other carriers here, that means bringing up services on multiple existing and new bands that range from 700 MHz to 2.6 GHz.
Seizo Onoe, vice president of R&D at Japan’s NTT Docomo, said the good news is carriers have been getting more spectrum from regulators. In addition, they can run 4G services in existing 3G bands, saving cost by reusing towers and RF chip sets in existing handsets. “We don’t need to cry over LTE spectrum fragmentation,” he said.
Nevertheless, he called for RFIC vendors to add more ports to their products and asked power amplifier makers to support a wider set of frequencies.
Mobile data traffic has far outpaced 2002 forecasts used for an assessment of spectrum needs at WRC (top). The U.S. FCC now forecasts a spectrum shortfall (middle). Meanwhile spectrum use for LTE is embracing about a dozen frequency bands (bottom).
Should we reuse the old spectrum using newer technologies after some a certain period of time?? ie. now 2G/3G/4G devices co exists. and within a few years 5G will be coming. If at any point of time, we decide we will stop using/making only 2G devices anymore, then after a few years(say 5), we can reuse the spectrum using newer technologies. The 5 years will be enough time for everyone to dispose their legacy devices :)
Dear markhahn there was a peer to peer group called the Poisoned Project that did something like that, - they hooked up all their members so each one having a particular song contributed proportionally the data for that song to a downloader, - fantastic download times, but unfortunately sabotaged by a deep packet inspection program from the rich boys called Sandmine.
Whatever your own feelings about that, from another side the Broadband Rollout in Australia, with it's huge capacities, may also be a way for overcoming the frequencies jam?
it should be obvious that treating spectrum as property is just wrong, since unlike, say, a toll road, there's no technical reason a device can't switch among 100 different providers and/or frequencies in the space of a minute. suppose phones simply conducted a reverse action when they needed service. (let the government tax the transaction, rather than obtain an inherently less efficient one-time payment for spectrum.) in other words, whoever uses the spectrum pays for it. can you imagine how customer-friendly this would be? it wouldn't necessarily preclude long-term service plans, but would inherently lessen the reliance on lockin. it would also improve coverage, literally exponentially, since your chances of getting service from _someone_ would the product of the probability of single-carrier coverage. this would also be much friendlier to the internet-of-things concept, since devices could obtain coverage without significant overheads. it would also create vast and numerous incentives for coverage where the big carriers do a poor job - perhaps there would even be arbitrage-like enterprises that simply provided patchwork-like coverage in the weak spots of bigger carriers. absurdities like arguing about micro-sim formats would disappear, since identities would simply be tied to billing, rather than requiring a separately maintained uniqueid.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.