|
|
US Builds Wired Homes
on PhonelinesHome network standardization is taking two approaches: HomePNA using telephone lines, and HomeRF using the 2.4GHz radio frequency band.
“The personal computers and peripheral equipment in the home today are going to be linked into a network,” predicts Dan Sweeney, manager, Business Unit, Home Networking Operation, Intel Corp of the US. “And that alone will create a home market worth a billion (US) dollars annually by 2003.” Companies are nurturing the home network market in the hope of capturing a dominant position.
In 1998, Intel, Advanced Micro Devices, Inc (AMD) of the US and other firms, together with Microsoft Corp and Compaq Computer Corp, both of the US, launched home networking standardization bodies: Home Radio Frequency (HomeRF) and Home Phoneline Networking Alliance (HomePNA). Shipment of integrated circuits (IC) for home networks has already started. From late 1998 and through 1999, a variety of home-use PCs are to be shipped with home networking functions included as standard equipment.
Aiming at Wired Home
The ultimate goal of the home network is the wired home, where all types of consumer electronics products are networked. For Japanese manufacturers, with extensive experience in consumer electronics, this represents an opportunity to enter the home network market. But they have been rather slow to react.
The aggressive moves of US manufacturers are due to two key differences between the markets of Japan and the US. One is the degree to which PCs and the Internet are used, and the other has to do with the portfolio of datacomm services to the home.
There is a dramatic difference between Japan and the US in the percentage of households which use PCs and the Internet. In the US, as of the end of 1997, 45% of households used PCs and 20% used the Internet, while for Japan these figures were just 25.2% and 12.1%, respectively, as of March 1998.
US manufacturers plan to construct a home network by connecting existing PCs in the home. If the business is developed from a core of existing PCs, there is a benefit in that: “The home network is already in the making, with printers and Internet access being shared,” says Yosuke Yukihira, assistant manager, Corporate R&D Strategy of Sony Corp of Japan.
This situation is similar to the appearance of local area networks (LAN), but there are supporting factors now that did not exist then. As Cyrus Namazi, chairman of HomePNA and product marketing manager, Network Products Division, AMD, comments, “One point of interest regarding the home network that did not exist for the LAN is that all the PCs in the home will be able to access the Internet.”
No New Wiring
“Just plug it and you’re ready to go,” say US home network PC manufacturers. As more homes come to own multiple PCs, new requirements are appearing (Fig 1).
When it comes to the data transmission path, manufacturers have demanded a network that doesn’t require new wiring. Wired options, therefore, consist of telephone lines (twisted pair), electric power lines, and television lines (coaxial cable).
Of these, most US manufacturers have selected telephone lines. “The biggest advantage of phonelines is that they are pervasive,” says Intel’s Sweeney. “The average US home already has four or five modular jacks.” Television cabling does not run to as many rooms as phonelines.
Telephone lines are superior when it comes to transmission speed as well: 1 Mbit/s for phonelines, versus about 350 kbits/s for power lines.
Even so, phonelines are not the answer for home networks in all homes. In these cases, the next best choice is a wireless system. Many manufacturers believe that a wireless system is the best option outside the US market, because the use of telephone or power lines for datacomm would require a match between a wide variety of national regulations.
HomePNA, HomeRF
The standardization organizations for phonelines (HomePNA) and wireless (HomeRF) already have over 100 member firms each. A large number of these members, including companies like Intel and Compaq, are members of both.
One distinct difference between the two organizations is the way in which they manage technology licensing (Fig 2). Licensing is a key issue in determining whether or not a standard will achieve widespread adoption.
Semiconductor chip manufacturers will require a license to develop home networking products compliant with HomePNA specifications. Manufacturers purchasing these integrated circuits (IC) to develop other products are not required to enter into a licensing agreement, or pay licensing fees. Cyrus Namazi, HomePNA chairman and product marketing manager of AMD Network Products Division explains: “We’ve made it possible for firms to manufacture products even without becoming HomePNA members.”
The HomeRF model, on the other hand, requires that firms first become members of the organization. Once a firm becomes a member it can utilize the HomeRF standard free of charge, with no demand for a license fee. If proprietary technology is to be included in the HomeRF specification, however, in principle the manufacturer has to provide it with free licensing. “License-free has been our guiding principle all along,” says Ben Manny, HomeRF chairman and engineering manager, Residential Networking, Intel Architecture Labs.
Flurry of Chips
There have been a number of announcements for home network ICs that use telephone lines (Table 1), all of which are compliant with the HomePNA standard – ensuring interconnectivity.
One of the reasons why companies have been so eager to jump on the HomePNA bandwagon is that there is no problem with conflicting specifications. In the past, it has been commonplace for multiple standards to arise for network IC applications, where two standards have competed for industry acceptance. One recent example has been the 56-kbps modem. “We learned a very painful lesson about split standards in the fight between K56flex and x2. Never again,” promises Jim Muth, product line manager, PC LAN Products, Rockwell Semiconductor Systems, Inc of the US.
None of the IC products being released are for dedicated home networking, but are instead 56k modem or Ethernet interface ICs that have added home networking functions. Manufacturers do not plan to produce dedicated home networking ICs in the future, either, but rather will use this function as a means of providing value-added content to existing 56k modem and Ethernet chips. Kevan Smith, director of marketing, Networking Division, National Semiconductor Corp of the US, reveals: “Home networking functions are scheduled to be implemented in all our Ethernet ICs.”
This strategy is possible because the HomeRun transmission scheme that complies to the HomePNA specification is not really very complex. Home networking functions can be implemented in only a relatively small circuit (3,000 to 4,000 gates). The PCnet-Home home networking IC from AMD, for example, sells for US$14.92 each in lots of 10,000.
Pushing 10 Mbits/s
HomePNA is currently available as version 1.0, but v2.0 plans to boost the transmission rate from 1 Mbit/s to 10 Mbits/s, which will make it possible to handle audio and video distribution. The standard is expected to be completed in the second half of 1999.
The transmission scheme is under investigation now, and the system being developed by Epigram Inc of the US has potential. Another 10-Mbit/s technology is being developed by TUT Systems, Inc of the US, which developed HomeRun. One feature of HomeRun is that it can use existing subscriber line and asymmetric digital subscriber line (ADSL) services to achieve 1-Mbit/s datacomm. This higher data rate is realized because the system utilizes the higher frequencies not available with these services (Fig 3).
This high-frequency region, however, is congested with noise from a variety of sources: amateur radio stations, amplitude modulation (AM) radio, and home appliances using motors such as washing machines.
TUT Systems developed its own encoding scheme, called TM32, which utilizes time differences in the waveforms to avoid the effects of noise (Fig 3b). The difference in a time (&Mac198;t) between two waveforms is allocated five bits of data as a symbol. The symbol cycle is maintained at 200kHz, achieving a data rate of 1 Mbit/s.
Epigram has not disclosed its technology in detail, but Tony Zuccarino, vice president, Marketing, comments, “We will provide 10-Mbit/s throughput by only slightly expanding the bandwidth (5.5 to 9.5MHz) used by HomeRun. Processing can be handled by commercially available DSP chips.” The assured transmission distance, however, will drop from about 150m under v1.0, to about 30m. The firm is currently negotiating licensing with IC manufacturers, and expects the first IC products to appear in the first quarter of 1999.
HomePNA is also working on even faster specifications. “We don’t feel that 10 Mbits/s is the final goal. While we haven’t set specific numeric targets like 20 or 40 Mbits/s, we do plan on being able to support MPEG-2 video,” says HomePNA’s Namazi. Epigram is also said to be developing a 100-Mbit/s scheme.
Turmoil in Wireless
While HomePNA has restricted itself to a single implementation through the telephone line, wireless home networking is being implemented through a host of different solutions.
A major factor for this is because the standardization body, HomeRF, does not have the dominant influence that HomePNA enjoys. There seems to be little concrete development of products complying with the shared wireless access protocol (SWAP) standard drawn up by HomeRF, while a number of manufacturers have already shipped proprietary systems. The result has been turmoil (Table 2).
One of the reasons for this is the complexity of SWAP. “SWAP has just too many functions,” complains Michael Hudson, vice president and chief technical officer, Communications Division, Diamond Multimedia Systems, Inc of the US. “If we made it according to the specs, it would be priced absurdly high.”
Another problem is that it remains unclear how it will coexist with Bluetooth, the wireless communication protocol spearheaded by Ericsson Telecom AB of Sweden. Bluetooth is rapidly gaining industry support, having about 200 corporate members as of the end of October 1998. Version 0.7 of the specification is being worked on now, and v1.0 is expected to be released in the first half of 1999. A software development kit (SDK) will also be made available.
Yasunori Akenaga, senior staff planner, Platform Architecture, Open Architecture, IBM Japan Ltd, says: “Even though HomeRF is aiming to provide home networking outside the scope of Bluetooth, both have often appeared as a group. Perhaps this is due to the fact that they share many points in common, like the 2.4GHz frequency band and frequency-hopping spread-spectrum technology.”
Price, Reliability, Speed
Compared to the wired solution, the wireless home network has a number of disadvantages: (1) its high price of several hundred US dollars, (2) problems in data transmission reliability, and (3) low transmission speed.
WebGear Inc of the US is working hard on the problem of price. By reducing the cost of its radio frequency (RF) module, the firm has managed to cut the price per transceiver to about US$85. While most manufacturers in the field use 2.4GHz band spread-spectrum technology, WebGear instead adopted a narrow-band modulation technology in the 900MHz waveband. “We just adopted the wireless module now used in cordless telephones,” revealed chief technical officer and chairman of the board, Mike Evans. Speed, however, has dropped to 500 kbits/s.
Reliability, the second problem, is the selling point used by Butterfly Communications, Inc of the US. As director of sales, Brian Ebert, points out, “By shortening the frequency-hopping interval to only 800ns, we assure transmission quality on a par with wired systems.”
The third problem of speed is being investigated by HomeRF on two fronts. Chairman Manny says, “Of these, the first is boosting speed with the existing 2.4GHz band. Tightening up the frequency-hopping interval will make it possible to improve throughput from the current rate of 2 Mbits/s to about 5 Mbits/s, and as high as 7 Mbits/s if we switch to direct sequence spread-spectrum. The other is to switch to the 5GHz band. We haven’t defined speed targets yet.” NTT Corp of Japan and Lucent Technologies, Inc of the US have already jointly proposed a 30-Mbit/s transmission scheme in the 5GHz waveband to the 802.11 Committee of the Institute of Electrical and Electronics Engineers, Inc (IEEE) of the US.
Consumer Control Market
Japanese manufacturers are entering home networking via so-called consumer electronics controllers, which control an array of home electronics equipment through a network. “As a consumer electronics manufacturer, we plan to get into the home networking market through consumer electronics controllers. Our strategy is not the same as a PC maker,” says Hidetaka Mori, manager, CE Media Business Headquarters, CE Media Development Center, Multimedia Systems Department, Sanyo Electric Co, Ltd of Japan.
The ideal solution would be the IEEE1394 network, which is capable of handling video streams. Japanese firms are continuing work on IEEE1394 networks as they start to enter the market for consumer electronics controllers.
Sanyo Electric is approaching home networks as part of its system IC business. The special Home Networking LSI Project team was formed in September 1998, and is now developing ICs specifically for consumer electronics control applications.
Sanyo Electric intends to provide a total control system capable of performing applications like home parties and home theater. Until lately, they have concentrated just on show demonstrations, but are now moving on to business trade negotiations.
Sharp Corp of Japan positions the two-way remote controller as the core of its home network, using it as a means to increase the value-added content of digital set-top boxes (STB) and other digital consumer electronics.
Wide Product Range
The appliance control market will take off in the US first. Already a wide variety of products are available, ranging from simple kits to control illumination via power lines, to complex systems delivering a home theater environment with the touch of a button.
Products complying with the X-10 standard for centralized control via power lines seem to be especially popular, with shipments of over a million units per month.
Japan does not have any appliance controller market yet, at least not to the extent that is enjoyed in the US, and as a result manufacturers are searching for different approaches. One is the power-control of appliances to reduce energy consumption. Matsushita Electric Industrial Co, Ltd of Japan, for example, is working on power-control for appliances. In cooperation with Toshiba Corp, Hitachi Ltd and Mitsubishi Electric Corp, all of Japan, it is developing a standard for appliance-control via electric power lines.
Consumer Standards Emerge
Two standards have emerged designed specifically for consumer electronics, the first of which is the Home Application Programming Interface (HAPI). This API provides PC-based centralized control of networked consumer electronics. The specification is being drawn up by the HomeAPI group, which consists of six firms: Compaq, Honeywell Inc of the US, Intel, Microsoft, Mitsubishi Electric, and Philips Electronics NV of the Netherlands. The public disclosure is scheduled for spring 1999, and Microsoft plans to ship a software development kit for Windows in the first half of 1998.
The other standard is IrDA Control (Fig 4), which is the latest version of the Infrared Data Association (IrDA) datacomm standard that uses infrared beams. It is intended for applications such as two-way remote controllers. A joint proposal by Hewlett-Packard Co of the US, Intel, Microsoft and Sharp, under the name IrBus, was formally standardized in February 1998.
Japan Makers Play Part
Both HAPI and IrDA Control are aimed at consumer electronics control applications. One key difference from similar standards that have evolved in the past is that Japanese manufacturers have been involved in the standardization meetings from the initial stages: Mitsubishi Electric in regard to HAPI, and Sharp over IrDA Control.
“They contacted us as a consumer electronics maker, because of our market presence in rear-projection television receivers and similar equipment,” says Naoki Yamauchi, department manager, Home Network LSI Development, System LSI Development Center, Mitsubishi Electric.
“Except for us,” says Hiroshi Uno, senior technical specialist, Software Research Laboratory of Sharp, “everyone involved in drawing up IrDA Control was a PC maker. The initial target for IrDA Control is PC peripherals, and the secondary target is consumer electronics, according to the initial specification.”
HAPI has three major aims. The first is centralize control with a PC, so that control of multiple pieces of equipment can be completed in a single operation. The PC is able to ensure interconnectivity in a multivendor environment, and adapt to changes in the system configuration caused by the addition or removal of equipment.
Secondly, the dedicated controller, making up the majority of consumer electronics controllers, should be replaced by a PC, if at all possible. This will minimize system cost.
Thirdly, a standard API is needed to create a market for consumer electronics control application software (Fig 5). This will promote participation by third-party vendors.
Interconnectivity is Key
Neither HAPI nor IrDA Control can construct a consumer electronics control system by themselves. They must be used in combination with other systems. Interconnectivity, including related home networking standards, seems likely to become a key issue in the near future (Fig 6).
by Masahiro Katoh
References:
- AMD
Butterfly
Compaq
Diamond
Epigram
Ericsson
Hewlett-Packard
Hitachi
HomePNA
HomeRF
Honeywell
IBM Japan
Intel
Lucent
Matsushita Electric
Microsoft
Mitsubishi Electric
National Semiconductor
NTT
Philips
Rockwell
Sanyo
Sharp
Sony
Sun Microsystems
Toshiba
TUT Systems
WebGear
