Qualcomm has a reputation for owning technologies and bitterly defending them in patent wars - a reputation it wants to dispel. Sanjay Jha, the chief operating officer of Qualcomm CDMA Technologies Group, Qualcomm's 3G silicon division, sat down with us and other journalists recently in San Diego, to lay out the company's strategy for wireless communications.
"The wireless world is changing," says Jha. "Globally, you can get 200 kbit/s to 500 kbit/s wherever you go. Five years ago, only 40 percent of the US had 30 kbit/s."
He's sold on mobile. "Broadband at home is one thing. Broadband wherever you go is a whole new thing. Windows, the IP protocol, the browser and the GigaHertz process all came together and an entity called the Internet emerged. Now, there's an opportunity for the wireless Internet to have a bigger impact than the wireline Internet."
"We want pocketable computing," he says, and he's keen to show us Snapdragon, a mobile chip Qualcomm introduced in November 2006. It combines HSPA mobile data GPS (global positioning). GSM,. EDGE and other standards . Based on the ARM, it operates at a GHz and on 500mW of power.
A patent powerhouse
Qualcomm made its name in CDMA, the current mobile technology in much of the non-GSM world. CDMA is still "alive and growing", says Jha (with some 380 million users, albeit including the 3G technologies wideband CDMA and CDMA2000).
But Qualcomm has its fingers in other standards, often with demands for licence fees for use of its patented technologies (for instance a suit against Nokia).
The process of extracting value from patents is far-reaching, even extending to technology which Qualcomm opposes, such as WiMax. "It is not possible to do mobile WiMax without using some essential patents form Qualcomm," says Jah, explaining Qualcomm's bid to extract licence fees from WiMax.
Why Qualcomm thinks WiMax will fail
He's not a fan though. "We're not believers in WiMax," says Jha, claiming that 802.16e "mobile" WiMax, "will not be truly mobile," suffering from latency in the handoffs and poor control. "It's not well designed, with a MAC derived from the DOCSIS cable standard," he says. "The IEEE recognize the fundamental issues - that's why they are working on the 802.16m standard [which promises 100Mbit/s mobile access]."
He reckons there will only be 25 million WiMax devices in 2012, compared to 600 million 3G devices. "WiMax won't have the scale or penetration." Even if it's in every Centrino laptop, that's only about 80 million devices, compared to a billion cellphones, he says. He dismisses the idea that the "open" nature of WiMax can make it fundamentally cheaper than 3G, given these scales.
"WiMax is not Wi-Fi," he says. "There will be the cost of spectrum," WiMax operates in licensed spectrum, while Wi-Fi is in unlicensed, he points out.
Ultra Mobile Broadband - better than WiMax?
Naturally, a better options is the more mobile-centric technology Qualcomm is backing, the Flash-OFDM it acquired with Flarion, which has become Ultra Mobile Broadband (UMB - also previously known as EV DO Rev C). This is a proposed next step for the CDMA2000 3G standard, has been approved by the 3GPP standards group - and is a trademark of Qualcomm.
It's also being proposed for the troubled IEEE 802.20 standards group, which has resumed work since the IEEE suspended it for three months last year.
Fixed mobile convergence - by multiple means
For a company that works with mobile operators, Qualcomm has a positive view of fixed mobile convergence (FMC), but, as Jha says, "It means so many things." The company is considering femtocells, small indoor base stations that use licensed spectrum to provide indoor voice - and 3G data - coverage, using ADSL on the phone line for backhaul. "As frequency goes up, coverage in buildings goes down," he says. A small local base station would allow a peak data rate of 20 to 25 Mbit/s, while allowing mobile carriers to offload the demand onto the wireline network.
At the same time, Qualcomm bought Wi-Fi chipmaker Airgo last year to provide technology for dual-mode handsets, which can hand off calls to indoor unlicensed Wi-Fi networks.
Airgo developed ways of doing MIMO which will be strong," says Jha. "With Airgo technology, the data rate decreases slowly as range increases." Multiple input multiple output (MIMO) is a way of running multiple Wi-Fi channels through the same air, which is the basis of 802.11n.
Qualcomm is in no hurry though. It plans to embed Wi-Fi in its chips by the end of 2008. "depending on uptake and usage models," according to Qualcomm's vice president of strategic products, Mike Concannon.
Meanwhile, Airgo is due to deliver fast 802.11n Wi-Fi, and Qualcomm will continue with Airgo's plans for standalone Wi-Fi chips for access points and devices, says Jha. The company will have a 2x3 (two transmit antennas, three receive antennas) 802.11n chip sampling in June.
A technology gambler
Overall, the licence fees it claims are the payback on bets Qualcomm makes on technologies, says Jha. "The longer the bet we make, the more it pays off." He's talking about OFDMA, the multiple-access version of orthogonal frequency division modulation, a term also claimed by Qualcomm's rival Runcom for its own technology.
As a "fabless" chip company that makes its money from licensing designs to other companies, Qualcomm has to protect its intellectual property, says Jha. The company looks after employees, and gives a bounty of "a few thousand dollars" to anyone who gets a patent, he says.