Receive new posts as email.
This site operates as an independent editorial operation. Advertising, sponsorships, and other non-editorial materials represent the opinions and messages of their respective origins, and not of the site operator or JiWire, Inc.
Entire site and all contents except otherwise noted © Copyright 2001-2006 by Glenn Fleishman. Some images ©2006 Jupiterimages Corporation. All rights reserved. Please contact us for reprint rights. Linking is, of course, free and encouraged.
A U.S. District judge has provided Broadcom with broad wins: The judge provided a permanent injunction against Qualcomm’s ability to sell cellular data chips that were deemed to infringe on Broadcom’s patents. Qualcomm must immediately stop selling WCDMA chips—that’s the technology used in UMTS and HSPA, 3G extensions of GSM—and can continue to sell EVDO chips through Jan. 2009 to customers they were selling products to as of May 29, when the infringement case was decided by jury. Qualcomm faces a mandatory set of royalties for those continued chip sales, and must negotiate with Broadcom for one set of royalties that the judge didn’t proscribe.
Qualcomm will be able to design around this injunction by producing new chips that don’t use the same processes that were found to violate Broadcom’s patents, but it will take a number of months, and leaves them out of the rising HSPA market that AT&T and T-Mobile will be pushing harder on as AT&T completes its faster 3G network and T-Mobile starts its real 3G buildout using frequency purchase at auction several months ago.
Big move for Broadcom, despite modest appearance: EDGE may be old technology, but it’s being rolled out like crazy in devices, since the networks have been built and are underused. Hence, the iPhone. Broadcom said today that Nokia, the dominant handset maker, will buy its EDGE chips from them. Broadcom says EDGE will be in 400m devices by 2009.
Broadcom is riding high on the various bits of news and legal decisions related to its ongoing patent battles with Qualcomm, including the decision by the US Trade Representative to allow a trade commission standing banning new handset models with Qualcomm 3G chips from entering the country; a judge’s slapdown of Qualcomm’s attorneys and standards’ group practices that have led the judge to find two patents invalid; and the potential for more of the same to come.
NTT DoCoMo is using MIMO for its so-called 3.9G network tests: The Super 3G networks would supercede HSxPA (high-speed downlink or uplink packet access) networks with potentially 300 Mbps of downstream performance. Today’s HSDPA service caps at 7.2 Mbps for a channel.
China will accept W-CDMA and CDMA 2000 as standards on its future 3G networks: These two international standards can be considered alongside the domestic TD-SCDMA technology China has been testing for years, and which the country issued a spec for in January 2006. Adding these options as acceptable 3G technology choices may simply be a sop to conform to international pressure or requirements.
China prefers to have unique, domestic standards, and to require foreign manufacturers to partner with Chinese firms to manufacture gear using non-domestic standards. It also likes to have specific methods by which the government can easily access information and data on Internet, wireless, and cellular networks, and their homegrown standard may be more amenable to that than those created for international markets.
A band of makers and operators will work on Long Term Evolution (LTE) wireless networks: LTE will ultimately supercede HSPA (HSDPA/HSUPA) standards, and will compete with mobile WiMax for newer markets. The new consortium includes major European operators (Orange, T-Mobile International, Vodafone) and major equipment makers (Alcatel-Lucent, Ericsson, Nokia, Nortel, Siemens). The trial will look at topping 100 Mbps.
Ars Technica writes up the coming specs for EVDO Rev. B: We won’t see it until for a couple of years at least—EVDO Rev. A is just really rolling out right now on Verizon and Sprint Nextel’s networks—but Rev. B will scorch Rev. A speeds. Rev. A might top 3 Mbps in ideal circumstances, but Rev. B showed 9.3 Mbps on average using a 5 MHz channel (3xEVDO, I believe, versus 1xEVDO today) in testing.
A fact that doesn’t get the same kind of attention as these speeds is how the backbone serves the cell sites that produce such high numbers. I have been told many times in the past by industry insiders that the U.S. cellular network only has a subset of its site served by more than T-1 lines or the equivalent. This is changing, but it’s a huge cost, and many locations can’t easily support fiber lines or faster service. This is where fixed WiMax and other wireless backhaul may come into effect, using licensed bandwidth on a point-to-multipoint basis to drive data to the cell sites which then distribute in a cloud around themselves.
Because of the slow pace of broadband speed improvements in the U.S., it’s likely that Rev. B could outstrip wireline broadband in many parts of the country and parts of many cities. While fiber to the home or node will be widespread by 2009, it’s predicted to be available to perhaps a quarter of the US population, of which only a portion will subscribe. Thus Sprint and its mobile WiMax and Verizon and its Rev. B could give wireline a run for its money. Except that Verizon’s current position is that cellular data isn’t a replacement for wireline; Sprint’s mobile WiMax rollout has a very different attitude, closer to Clearwire’s.
The Wall St Journal looks into wireless crystal ball: The newspaper predicts substantially more mobile video as services like Qualcomm’s MediaFlo roll out. MediaFlo uses spectrum Qualcomm won in the former UHF TV band that will allow them to broadcast high-quality video and audio to cell receivers without the compromises required to share cellular voice/data bands. Verizon is offering MediaFlo in 20 markets for $15 per month by itself.
MobiTV will continue to expand its offering; they offer television over Wi-Fi and broadband networks. They hope to offer a converged plan in which the same programming will be available for a single fee across many different networks. MobiTV isn’t a replacement for IPTV, which requires high home bandwidth; rather, it’s a streaming TV on PC/handheld offering—for now.
The article also notes that Wi-Fi is being built into an increasingly large number of cell phones—80 so far—and the need to hook these phones into hotspot networks becomes ever greater.
Intel will ship the next Centrino platform without an option for the worldwide HSDPA 3G cell standard: The partnership with Nokia was supposed to lead to both Wi-Fi and HSDPA (GSM’s top-speed 3G offering) in a single package. Instead, when Santa Rosa launches in the second quarter of this year, Wi-Fi will ride solo. This is the end of a strategy that dates back to the first Centrino launch when an Intel executive told me that it was inevitable that Wi-Fi and cell would both find their way into the platform. Four years later, no such luck.
The bottom line appears to be dollars. Intel told InfoWorld that laptop makers don’t want to pay to have HSDPA integrated in every machine. Intel will consider adding 3G in future platforms, however. Specific manufacturers have chosen to include EVDO and HSDPA embedded in their laptops, but that increases cost and requires consumers or businesses to order specific models.
The upshot is that the cell networks lost the chance to avoid subsidizing hundreds of thousands or millions more PC Cards and ExpressCards, and gain a new audience that could flip a switch to gain cell data service.
Ultra Mobile Broadband will apparently supercede EVDO’s name and standards: The CDMA Development Group, a trade association, is looking to the simpler name to brand faster speeds coming in future standards. While EVDO Rev. A is rolling out now with much higher rates and Rev. B is on the horizon with about 50 Mbps of downstream peak service, Rev. C will be rebranded as UMB—and achieve 280 Mbps downstream in about 2009. UMB will use both MIMO (multiple receive/transmit antennas for beamforming and multiple spatial streams) and SDMA, which allows datastreams to be steered to particular clients, reusing the same frequencies over space.
This BetaNews article also notes that there are 44.4m EVDO subscribers worldwide and 83.6m UMTS users. 1xRTT + EVDO totals 267.2m, however.
T-Mobile will use gear from Nokia and Ericsson for its U.S. rollout of true 3G cell: T-Mobile’s American air said it will spend about $2.7b over two to three years to upgrade its network to 3G using the new spectrum it acquired at auction this fall.
T-Mobile said today that they would use their new spectrum licenses to build a 3G network: The company spent $4.2b to acquire about 23 MHz of nationwide spectrum in the 1700 MHz and 2100 MHz bands. They will deploy what’s being called UMTS across the US starting later this year; they have equipment deployed already, waiting to be turned on, apparently.
UMTS is a generic term for 3G GSM-evolved networks, but is also often used to refer to the first, slow flavor, somewhere between EDGE/1xRTT speeds and EVDO. HSDPA, which is what T-Mobile will be largely deploying (although details are a little scanty), operates around EVDO Rev. 0 speeds at present. Near-term versions of HSDPA should match EVDO Rev. A. GigaOm writes that the company is discussing the network as HSDPA-ready rather than as running HSDPA.
The frequencies that T-Mobile’s UMTS service will work on don’t match up worldwide, requiring new equipment. Europeans do offer UMTS over 2100 MHz, but not in the same configuration that T-Mobile will deploy.
T-Mobile will rev up the service in limited markets this year, with the full deployment taking through 2008, apparently. Handsets and services won’t be launched until mid-2007. The company also said that UMA (unlicensed mobile access), which allows seamless roaming between Wi-Fi and cell networks while making voice calls, will be in trials this year.
eWeek is saying that T-Mobile will announce its plans for a UMTS network on Oct. 6: The report says that T-Mobile’s recent successful cash blowout on spectrum in the FCC’s advanced wireless auction will lead to an 18-month rollout of UMTS service using 1700 and 2100 MHz, which are not used elsewhere for UMTS. Which means that existing UMTS devices will not work on T-Mobile’s network; it will need to have custom equipment made (not a big deal these days), and hope for multi-band cards that will work worldwide, including on its network.
UMTS describes both the scope of 3G network types in Europe, as well as a particular lower-speed technology implementation as rolled out worldwide. HSDPA, UMTS’s successor, is comparable in its current version to EVDO Rev. 0 (about 50 to 150 Kbps upstream, and 400 to 700 Kbps downstream, in typical performance). UMTS runs perhaps 50 to 100 Kpbs upstream and no more than 400 Kbps downstream.
Silicon giant Intel partners with Nokia to put cell data modems on motherboards: Nokia will manufacture HSDPA modems that will be embedded in Santa Rosa motherboards, Intel said last week, which is the next generation of Centrino. Santa Rosa will encompass 802.11n, the Core 2 Duo processor, improved graphics performance, and other upgrades; it will ship in the first half of 2007.
Everyone knew the day would come: Verizon is testing Nortel ‘s EVDO Rev. A equipment, which would be a relatively simple upgrade from the Rev. 0 gear that they and Sprint Nextel use today. Rev. A dramatically boosts total and expected individual user uplink rates from what is often 50 to 150 Kbps today to a raw rate of 1.8 Mbps that’s expected to offer 150 to 250 Kbps on a typical basis to each user. Downlink speeds jump, too, from about 2 Mbps to 3.1 Mbps peak rate, which should translate to 500 to 800 Kbps for an individual user routinely.
Sprint had said they would start the move to Rev. A next year. Nortel stated that Verizon would begin the upgrade in third quarter of 2006! Yes, uplink speeds matter, and this is another clear sign of the consumer-as-producer economy, in which people creating—taking digital photographs, for one—need more upstream bandwidth.
Technology Review reports that NTT DoCoMo has a prototype wireless network operating at 2.5 gigabits per second (Gbps): It’s mobile, working at 20 kph, and could move a DVD across in about 10 seconds. They use MIMO with six antennas, but they mention neither spectrum band nor channel width. And it’s impossible to stick this receiver in a handheld device right now, either.
There’s an obscure note about tweaking QAM, part of the modulation used to carry multiple symbols over a single channel or subchannel, but increasing the number of symbols per modulation (the effective raw bit rate) typically requires a greater signal to noise ratio. That is, more power or more bandwidth than a channel that can only carry a smaller amount of symbols.
The article identifies this as part of the 4G direction, in which everything is Internet Protocol and speeds are orders of magnitude higher. But it cautions that this is several years away.
Qualcomm said at a briefing attended by Alan Reiter recently that they had bought the rights to UHF channel 55 for their MediaFLO service in the U.S.: The service will allow cell operators to offload push-media—streaming TV, audio, and targeted data—to spectrum other than that scarce and expensive stuff they use for voice calls. It also ties the operators directly to Qualcomm, which will control that spectrum, and which should make some carriers slightly nervous. No knock on Qualcomm, but any single vendor solution outside the control of an operator that combines hardware, software, patents, and spectrum should cause cold sweats.
Now, I linked to Alan’s May 31 report before, but he queried Qualcomm about how they’d get clearance to channel 55, given that the channel was still in use by some stations around the country. This is a quandary faced by all those companies that won currently occupied UHF spectrum for future data services.
Alan graciously is allowing me to write up what Qualcomm told him, as I prompted his questions. Qualcomm told him that while they won’t speak specifically about regions in which the channels are occupied—as they are negotiating to clear spectrum in each market—they have a deadline now set by Congress of Feb. 17, 2009, in which all broadcasts in those bands must halt.
Qualcomm did say that because it’s a matter of public record, they could point to agreements with specific stations, including WLNY (Riverhead, New York), WACX (Leesburg, Florida), and KWDK (Seattle-Tacoma, Wash.), among others.
They’re planning to launch access after 2006 regardless of whether they have the entire country cleared, and regardless of some action they’re looking for on interference from adjacent UHF channels that are still in operation in some markets (54 and 56).
Alan Reiter writes up a briefing on Qualcomm’s broadcast format for cell phones: MediaFLO will use former UHF Channel 55 in the U.S., carrying 20 video and 10 audio channels over 6 MHz. It’ll also stream 800 minutes of short, downloadable clips across each day and provide “IP datacasting”—data broadcast to a granular subset of users, such as stock ticker or sports updates as games progress. Verizon has committed to the service.
Qualcomm may have adapters ready for EVDO Rev. B in late 2007: The current deployed CDMA2000 1xEVDO Rev. 0 (zero) standard runs at about 400 to 700 Kbps (rated) downstream and 50 to 70 Kbps up. The Rev. A improvement increases downlink speeds by 30 percent and could double uploads. This starts to move EVDO into the DSL range, and ubiquitously. (There are still issues of latency, operator limits on service and bandwidth, and other factors, of course.)
Rev. B, Qualcomm said last week according to News.com, will further increase speeds, offering download rates of 1.3 to 2.4 Mbps and upload rates of 210 to 432 Kbps. Now we’re talking something approaching real broadband. The PC cards and stand-alone modems to use Rev. B may be available in late 2007.
While the article maintains “The increase in speed puts wireless broadband on equal footing with DSL services, which offer similar speeds” that’s only with a moderate to slow version of today’s DSL. Cable modem speeds typically start much higher than Rev. B, with double to triple the upload speed as a starting point, while DSL, cable, fiber to the home, fiber to the node, broadband over powerline, fixed WiMax (16d and 16e), and even new flavors of metro-scale Wi-Fi based on 802.11n and MIMO technologies will deliver speeds 5 to 10 times faster than Rev. B by the end of 2007, passing by at least tens of millions of homes.
Of course, with EVDO, Rev. B transceivers could reach hundreds of millions of potential users, making it an interesting option. The notion that standalone modems will be part of the reference designs released for Rev. B means that the vision of Monet Wireless could finally be fulfilled on a large-scale: rural areas with fewer broadband access methods could turn to licensed 3G cellular as a means of having broadband. This is the same sub-market that 802.16e (fixed/nomadic/mobile) WiMax is after, and that may be where big battles are fought—not in the cities, but in the suburbs, exurbs, and countryside.
The VP of engineering operations at T-Mobile USA says 3G in 2006, maybe 2007: They’re testing in some markets, but they need to win some auctions next year for spectrum, it sounds like. The company won’t say which markets they have test setups in nor which flavor of 3G they’ll opt for. T-Mobile currently offers unlimited GPRS nationwide as a flat rate add-on to its voice subscribers. [link via TechDirt]
Qualcomm announced a host of future additions to the EVDO and HSDPA standards: EVDO Rev B, mentioned in the previous post, is just one of a list of DMMX (DO Multicarrier Multilink Extensions) and HMMX (HSDPA MMX) add-ons. The multicarrier, multilink means that both standards will be able to work over protocols and bands simultaneously instead of requiring all service in a single band on a single carrier.
The EVDO Rev B speed boost can be accomplished by bonding 1.25 MHz channels, the current 1x channel width, in agglomerations of up 15 or 20 MHz total which would allow 73.5 Mbps downlinks. Even a single 1.25 MHz channel will increase from 3.1 Mbps with Rev A to 4.9 Mbps with Rev B.
They’ll also support some tweaky radio frequency and antenna additions that should increase range at lower signal levels; a new codec will improve voice quality; and GPS, Wi-Fi, and Bluetooth will all be integrated into their offerings as well as OFDM as an encoding method for streaming audio and video.
Their goal is a multi-tasking convergence that allows simultaneous multi-band, multi-function operation. Thus GPS tracking, Internet mapping software, and a phone call could happen at once, or VoIP and Web page viewing with groupware functions.
It’s a brave new world, and Qualcomm has just leapt in with techniques and standards that show their extend, embrace, and conquer attitude.
The revision to the EVDO standard might be available within two years: The technology binds together many more channels, which means it’s more spectrum-intensive on the operator, according to the details provided in this article. But in areas in which the frequencies are available, this is a huge jump from the 3.1 Mbps of Rev A, which should start rolling out this next year.
Somebody left the door open over at Cingular, showing UMTS deployment: A customer-service portal has been left open, discovered by the folks at HowardForums, showing Cingular’s plans to announce UMTS service Nov. 1, 2005. UMTS is a catchall term for 3G cellular data, and the HSDPA (high-speed downlink packet access) flavor that’s faster than the first-generation plain UMTS and CDMA-based EVDO is considered UMTS.
The open site (so far) shows 18 markets that will launch Nov. 1. The maps for these markets show the Nov. 1 launch plans and also the area planned for expansion by 2007. In the Seattle-Tacoma market, a pretty enormous area is covered by the first phase; the state capitol and surrounding areas are slated for phase 2. (I haven’t posted the maps as it’s likely there will be a take-down request from Cingular. The HowardForums pages have the maps, for now.) [link via Engadget]
Our fine colleague Nancy Gohring got ahold of this story and Cingular commented on the Web site: no comment! But they said they’re on track to deploy in 15 to 20 markets this year. Which sounds an awful lot like a comment. Nancy’s analysis shows that some markets may get plain old UMTS (200 to 300 Kbps) and others HSDPA (400 to 700 Kbps).
This article tries to make Nokia out to be savvy, but they sound a bit flighty: Nokia is focused its laser beam—everywhere. On 3G cellular, on Wi-Fi, and mobile WiMax. The company is right to do so, but it’s a little scattershot in this description of the growing data side of their handset and communicator business. Nokia has committed to including Wi-Fi on all new higher-end devices, but the article notes that some devices have “eight radios and antennas”—more likely bands and chips—to support GSM and 3G worldwide along with Bluetooth and Wi-Fi.
InfoWorld looks at how businesses are already adopting 3G flavors to solve problems for roaming employees: The various kinds of 3G here or on the way will bring higher speeds, as will mobile WiMax when it’s available in 2007 or later. But will mobile WiMax’s cost of deployment overwhelm its advantages? This article argues that ubiquity with mobile WiMax is too high a price to compete with cellular.
The division is safe: The company will seek more spectrum via auctions next year.
T-Mobile USA’s head says the company to offer 3G services in 2007: T-Mobile’s plans have lagged the other soon-to-be-three much larger U.S. competitors due primarily to a spectrum deficit, but also the desire to wait out early 3G technology, it seems. T-Mobile has unlimited GPRS plans, which seems a little pokey (though reasonably priced) contrasted against unlimited EDGE (Cingular, nationwide) and 1xEVDO (Verizon Wireless, certain major cities).
T-Mobile will have to buy spectrum, and believes this will be easier and cheaper with fewer competitors in the market. The company will start offering 3G in the second half of 2006, but won’t add applications like music downloads until 2007.