IN THE PIPELINE: Radio Free Intel To Connect Gadgets
Imagine a laptop that stays connected to the Internet no matter where you are, or a cellphone that picks up crystal clear reception anywhere in the world.
Intel Corp. (INTC) is working on a chip that would allow such connectivity.
Wireless devices today are like computers that can run only one program: they are hardwired to work in one standard only. Computers use Wi-Fi; cell phones use CDMA, GSM or any other number of platforms; wireless headsets and other devices use Bluetooth. Communication among various devices is complicated and spotty. A single device to incorporate standards would be bulky and expensive.
To solve this problem, Intel is using a technology called software defined radio to develop a platform smart enough to sniff out and adapt to the standards in any given environment. The company wants to put one such radio on all its chips, enabling virtually hundreds or thousands of devices to talk to one another. Radio is often thought of as only AM and FM, but all wireless transmission is radio, including TV, satellite and cell phones. Think of the implications when your refrigerator, cable box, computer, PDA, digital camera, and phone are all connected to the Internet and one another.
A chip with an integrated radio is also a step towards miniaturization. Radios are currently separate components, taking up extra space and power. With integrated chips, video conferencing capabilities could someday be incorporated into a wrist watch. A cell phone could be shrunk down to a keychain that's wirelessly tethered to an earbud.
Intel's idea, dubbed "Radio Free Intel," was put forth by chief technology officer Patrick Gelsinger more than a year ago at the company's annual developer forum. A leader among PC chip makers, Intel does not dominate the communications market, which is more the domain of Advanced Micro Devices Inc. (AMD), Texas Instruments Inc. (TXN) and Motorola Inc. (MOT).
Intel has plenty of competition in the arena of software defined radio. Texas Instruments and Motorola aren't going to give up their toeholds easily and smaller companies, such as Sandbridge, General Dynamics Corp. (GD), and Analog Devices Inc. (ADI), have been pursuing the technology for years.
Intel is "running to get in front of a parade that's already begun," said Will Strauss, president of Forward Concepts, a Tempe, Ariz., market research firm. "It's a lot of posturing."
But as a giant in the industry, Intel's interest in the technology is making a big splash and will likely spur development.
"It's good when big manufacturers get involved, it validates the things that the rest of us are doing," said Rupert Baines, vice president of marketing at PicoChip, a privately held communications equipment maker based in the U.K.
Considered the Holy Grail of radio technology, software defined radio is slowly being deployed in military and commercial use. Widespread consumer use is seen five to 10 years off.
"Right now, we're in kind of middle ground," said Craig Mathias of Farpoint Group, a consulting firm in Ashland, Mass.
General On A Payphone
Software defined radio started in 1983 with the U.S. invasion of Grenada. At that time, radios used by the various branches of the armed forces couldn't communicate with one another. The story goes that a general had to resort to using a payphone in order to talk to someone at another military unit.
To avoid more incidents like that, the military started looking into radio technology that could adapt to different frequencies. Telecommunications companies also became interested in the technology, initially as a way to improve security. Now more are starting to see it as a platform to make devices more flexible.
Software defined radio technology is currently being developed for cellphone towers and the military. There is also lots of interest in using the technology to improve public safety systems. Fire departments, the police and others, like the military, operate on different communications systems that don't work with one another.
PicoChip recently unveiled a software defined platform for cellphone towers. Towers today are dedicated to one specific standard, making the process of upgrading or changing standards time-consuming and costly.
"They call it the forklift upgrade," said Baines. "That's something you don't want to do very often."
Alterations to software defined radio, on the other hand, are relatively easy and inexpensive to implement. It does not require any change to hardware in the tower and can be performed from a central location. The technology also enables carriers to easily switch standards, such as from CDMA to w-CDMA, or to carry multiple standards.
According to Baines, using software defined radio can save wireless carriers tens of millions of dollars in development costs for each new market they want to enter and save tens of thousands in manufacturing costs for each new cellphone tower.
Meanwhile, the military, working with companies like General Dynamics Corp. (GD), has developed radios that can adapt to different frequencies. The Navy is currently using General Dynamics' digital modular radio, which is the size of a TV set, weighs about 200 pounds and costs just under $400,000. The units are designed to replace the radio room in a submarine, which typically has racks crammed with radio equipment, said Bill Rau, business development manager at the company. A single unit is capable of handling about 120 waveforms, compared with a typical radio, which can handle two waveforms at most.
General Dynamics is also rolling out a new software defined radio product that is more mobile. Weighing around two pounds and capable of handling five waveforms at a time, the unit makes it possible for soldiers to more easily communicate with various units without lugging around multiple radios, Rau said. But at roughly $900 to $14,000 a pop, the product still needs some major tweaking before it becomes acceptable for mainstream consumers.
"For handsets, there are three things that matter: cost, cost and cost," said Baines. "After that is battery life and miniaturization."
"The economics for handsets are slightly different. A telecommunications company might be prepared to spend 10% more on a basestation, but the consumer doesn't think that way about a handset. They just think it's a more expensive one and want to know if they would benefit from that extra cost," Baines explained. Manufacturers must either reduce the materials cost or deliver big benefits to the end user.
Duct-Taping Phones
For the near future at least, the cost of software defined radio could be greater than that of "duct-taping" several units together the way dual-band phones currently do. Dual band phones are essentially two phones that have been stuck together into one case - there is a lot of replication. Each segment of the phone is dedicated to one standard - either analog or digital. The phone's operation physically switches between one or the other.
Sometimes it's easier just to cobble three (standards) together, said Paul Kolodzy, director of the Center for Wireless Network Security at the Stevens Institute of Technology in Hoboken, N.J. This is especially true if a device has only to switch among a few standards.
Intel intends to add radios to its chips at no additional cost by making them in silicon, the same material it uses to make chips. Radios have traditionally been made of silicon germanium, which is more costly.
Pundits point out that as this hasn't been done before, there is no assurance radios can be made using this process. But Intel is counting on Moore's Law , which states that chips double their capacity every 18 months, to provide the extra processing power needed to run a radio.
As with any new technology, reaching a critical mass of users is an uphill battle, but once reached, widespread adoption is seen as almost inevitable.
Ben Manny, director of wireless technology development at Intel, made an analogy to USB ports. USB ports, which connect cameras, MP3 players and other devices to PCs, have become so ubiquitous that virtually all PCs ship with USB ports as a standard today.
"We see radios following a similar evolution," he said.
英特尔的伟大梦想
想象一台无论在哪里都能一直与互联网相连的笔记本电脑，或者一部在世界上任何一个角落都能清楚地接收信号的手机。
英特尔(Intel Corp., INTC)正在开发一种能够实现这一目标的晶片。
今天的无线装置就好像一台只能运行一个程序的电脑，它们只能在一种标准下运行。电脑使用Wi-Fi标准；手机使用CDMA、GSM或许多一些平台；无线耳机和其他设施则使用蓝牙。各种设施之间的通讯非常复杂且不稳定。如果要用一个装置合并所有这些标准，那将既笨拙又昂贵。
为解决这一难题，英特尔使用了一种名为无线电定义软件(software defined radio)的技术来开发一个平台，该平台的智能化能够达到相当水平、足以探测并适应任何环境下的标准。该公司希望在所有晶片上安置这样的无线电，从而使数百甚至数千个设备可以真正互相通讯。想象一下，当你的电冰箱、分线盒、电脑、个人数字助理、数码相机和电话都能和互联网相连并相互连接时，那将是一个怎样的情形。
集成无线电晶片还将走向小型化的一步。目前的无线电由单个零部件组成，占用额外空间并消耗额外能量。而在集成晶片下，视频会议功能有一天会整合在一只手表中。一部蜂窝电话可以像一个钥匙链那么小，通过无线连接到耳机上。
英特尔的这个策略名为"Radio Free Intel"，由首席技术长帕特里克?盖尔辛格(Patrick Gelsinger)在1年多以前的公司年度开发人员论坛上提出。尽管英特尔是个人电脑晶片制造商的领袖，但它并没有主导通讯市场，这个市场属于高级微设备公司(Advanced Micro Devices Inc., AMD)、德州仪器(Texas Instruments Inc., TXN)和摩托罗拉(Motorola Inc., MOT)的领地。
在软件无线电这个舞台上，英特尔的竞争阻力很大。德州仪器和摩托罗拉不会轻易拱手让出各自在市场中的地位，而小公司，如Sandbridge、General Dynamics Corp. (GD)和Analog Devices Inc. (ADI)数年来也一直在开发这种技术。 亚利桑那州市场研究公司Forward Concepts总裁威尔?斯特劳斯(Will Strauss)说，在这个已经开始的比赛中，英特尔正大踏步地前进，走在队伍的前面。英特尔已经做出了许多姿态。
但是作为行业巨头，英特尔对该技术的兴趣已引起了巨大轰动，可能会推动该技术的开发。
英国的私人通讯设备制造商PicoChip的营销副总裁鲁伯特?拜恩斯(Rupert Baines)说，大公司参与是好事情，这证明了其它公司正在从事的工作是有意义的。
软件无线电技术被认为是无线电技术的圣杯，该技术正在缓慢地应用于军事和商业用途，但预计广泛地应用于消费领域还要在5至10年之后。
马萨诸塞州咨询公司Farpoint Group的克雷格?马蒂阿斯(Craig Mathias)说，目前我们正处于该技术研究和应用的中间阶段。
- -软件无线电始于1983年美国入侵格林纳达之时。当时，各部队使用的无线电无法互相联络。据说，一位将军为了和另一支部队取得联系而不得不借助付费电话。
为了避免更多类似事件发生，军方开始研究能够适合不同频率的无线电技术。电信公司也开始对该技术产生了兴趣，最初是计划将该技术用来提高安全性。而现在，越来越多的人开始将该技术视作开发适应性更强的装置的一个平台。
目前，对软件无线电技术的开发是为手机信号发射塔和军队服务。也有很多人希望通过应用该技术来提高公共安全系统。和军队一样，消防队、警察局和其他部门的通讯系统也各不相同，不能互相兼容。 PicoChip近期推出了一个用于手机信号发射塔的软件无线电平台。目前的发射塔都只符合一个具体的标准，升级或改变标准都费时费力。
另一方面，对软件无线电的改变则非常容易而且成本很低，不需要改变发射塔的硬件，只在中央位置进行操作就可以了。该技术还可以使运营商轻松转换标准，比如从CDMA转换至w-CDMA，或者携带多个标准。
据拜恩斯介绍，在进入每一个新市场时，使用软件无线电可以为无线运营商节省数千万美元的开发成本，每一个新的手机发射塔可以节省数万美元的建造成本。 同时，与General Dynamics Corp. (GD)等公司合作的军队也已经开发出了能够适合不同频率的无线电。
美国海军目前正在使用General Dynamics的数字模块无线电，该产品相当于一台电视机大小，重约200磅，价格略低于40万美元。该公司的业务发展经理Bill Rau说，这些装置用来替换潜艇中通常每个架子上都堆满无线电设备的无线电室。每个装置能够处理大约120个波形，而相比之下，普通的无线电最多能处理两个波形。
General Dynamics还将推出一个更加便携的软件无线电产品。Rau说，该产品重大约2磅，能同时处理5个波形，使士兵可能更容易地与各个部门进行联络，而不必使用多个无线电来进行。但是该产品单价在大约900至14,000美元之间，在被主流消费者接受之前仍需要进行一些大的调整。
拜恩斯说，对手持设备来说，有三个影响因素：成本、成本还是成本。其次是电池寿命和小型化。
他解释说，手机的经济效益略有不同。一个电信公司可能不在乎在一个基站上多投资10%，但是消费者对手机的看法则不同，他们只想，这是个昂贵的东西，他们想知道是否多花的钱能给他们带来好处。生产商要么需要降低原材料成本，要么给最终用户很大的好处。
至少在短期内，软件无线电的成本将大于像目前的双波段电话那样将几个装置组合在一起的成本。双波段电话实际上是两个电话，它们被装入同一个外壳中，其中有许多功能重复。电话的每个部分要么符合模拟标准，要么符合数字标准。双波段电话的操作实际上是在两个电话之间转换。
新泽西州Stevens Institute of Technology的无线网络安全中心负责人保罗?克罗茨基(Paul Kolodzy)说，仅仅将三个标准组合到一起有时比较容易，尤其是在某一装置只需在几个标准之间转换的情况下。
英特尔计划通过以矽元素生产无线电，从而可以不花费额外成本将无线电补充在其晶片上面。矽是英特尔用来制造晶片的同一种材料。传统上无线电是用更昂贵的矽锗来制造。 权威人士指出，因为以前没有这样做过，不能确保可以用这种方法制造无线电。但是英特尔寄希望于公司创办人之一摩尔的定律(Moore's Law)，该定律称，为满足无线电运行所需的额外处理能量，晶片的数量每18个月会增长一倍。
和任何新技术一样，要达到至关重要的用户规模会越来越艰难，但一旦达到这个规模，大范围应用几乎是不可避免的。
