[Cover Story] Can Bluetooth Really Connect?
Version 1.0, the official Bluetooth specification, was released in July 1999. The first products are expected soon, but there is still a major obstacle: widespread concern over whether manufacturers and equipment can interconnect successfully?
Bluetooth-compliant products are expected to appear around the end of the year, or early in 2001 (Fig 1, Fig 2). The key point will be whether or not interconnection of different types of equipment and equipment from different vendors can be assured. If not, the interface could end in failure, as so many other "standard" interfaces have in the past.
The Bluetooth Special Interest Group (SIG) stressed assurance of interconnectability from the very start. Learning the lessons of the past, it has formulated three frameworks to make this possible.
First, it has provided a framework for correcting errors (bugs) and vague sections in the specification. Normally once a specification is issued, manufacturers begin implementing integrated circuits (IC) and firmware based on it, and when these early designs are interconnected, things rarely go well. This is because usually there are places in the specification which are interpreted differently by different engineers.
To achieve a common interpretation of these vague portions of the implementation spec, manufacturers must bring all of their "compliant" products together and evaluate them. For the Universal Serial Bus (USB), the USB implementers held a "PlugFest" to accomplish this. In IEEE1394 as well, the developers' conference held by the IEEE1394 Trade Association (TA), rather than a few key manufacturers, served the same purpose. Each firm ran its own interconnect tests, and found its own errors ( Table 1).
The Bluetooth SIG has already held two "UnPlugFest" events. On both occasions the vague portions of the middleware (such as descriptions dealing with RFCOMM and piconet) were revised.
Defining Implementation Methods
Second is specification of the middleware implementation method itself. In Bluetooth, these implementation methods are called profiles, and, as long as the protocol stack is made in accordance with the profile, interconnection with different equipment is assured through the application.
There was nothing like these profiles in the Infrared Data Association (IrDA) standard. In IrDA, physical layer interconnect was assured, but differences in middleware protocol stacks above the physical layer made multivendor interconnect impossible at one time.
Later, various protocols were developed to assure interconnectability, such as IrTran-P for still pictures and IrMC for mobile phones.
Bluetooth not only defines the profile, but also requires products to specify which profile(s) they comply with.
A user can immediately determine which profile is supported by any piece of equipment.
The third framework is the Qualification Program for Interconnectability. The testing verifies compliance with the Bluetooth specification as well as interconnect performance. Without this qualification, products are unable to affix the Bluetooth Logo demonstrating compliance. The goal is, of course, to protect the quality of the Bluetooth interface.
Other interfaces such as IEEE1394, IrDA and Ethernet do not offer this type of qualification program. It is used, however, in the Group Special Mobile (GSM) standard for mobile phones. It is likely that a similar system will also be adopted in USB 2.0, and the Op i.LINK optical interface for IEEE1394 being jointly developed by Sony Corp and Sharp Corp, both of Japan, and others.
In fact, it is possible that such a qualification program is becoming essential for interfaces.
Version 2.0, a Future Task
Standardization is continuing in an effort to accelerate product release to the market, with activities underway in parallel by the Bluetooth SIG and IEEE802.15 ( Fig 3).
The standardization schedule calls for Version 1.1, correcting various bugs in the current Version 1.0 issue, was scheduled to be released in the second half of August 2000.
The first official Bluetooth spec, Version 1.0, was issued in July last year. It was followed by Version 1.0B in December after a number of vaguely-written sections were revised.
The goal of Version 1.1, the second revision, will be to fix the remaining bugs and define the implementation method for piconets. In Version 1.0B there was no detailed description of piconet implementation, with the result that various implementations were developed, leading to a number of interconnect issues. A source at Silicon Wave, Inc of the US explains, "It can be processed in the Logical Link Control and Adaptation Protocol (L2CAP), or by the application software itself. The implementation was left up to the manufacturer."
In Version 1.1, the revisions made to the unclear text should vastly improve interconnection between IC over performance achieved with Version 1.0. And with the completion of Version 1.1, interconnect at the physical layer is essentially assured.
After public disclosure of Version 1.1, work will begin full-scale on Version 2.0, which will provide a faster data rate and support for new profiles. The Bluetooth SIG already has nine workgroups to investigate new profiles for possible inclusion in Version 2.0 ( Table 2).
Version 2.0 will not only provide support for new profiles, but will also change the radio frequency (RF) specification in two key areas: the peak data transfer rate will be boosted and interference with other radio communication schemes will be slashed. The peak data rate will be increased from the current 1 Mbit/s (effective 721 kbits/s) to between 2 and 12 Mbits/s. The same 2.4GHz waveband will be used, and the same spread-spectrum frequency hopping for data send and receive. The primary modulation, however, will probably be changed from 2-level Gaussian Frequency Shift Keying (GFSK) to 4-level.
Interference with other radio communication schemes essentially means interference with IEEE802.11b. The "WLAN co-existence" workgroup was formed to resolve this problem, and is expected to add improvements to the packet processing method and other areas.
The other key point of Version 2.0, new profiles, has also just begun. Most of the workgroups are now determining requirement specifications (MRD) for various target markets, and while they hope to complete work before the end of 2000, it may be difficult. It will probably be the second half of 2001 before the new profiles are all ready to go and the high-speed RF spec is solid. As a result, products complying with Version 1.1 will remain the most common for some time. "It is more important for us to create a product market with Version 1.1 right now, than to worry about Version 2.2," says a source at Toshiba Corp of Japan.
Standardization at IEEE802.15
In parallel with standardization activities by the Bluetooth SIG, the IEEE802.15 Committee is also advancing standardization. IEEE802.15 derives from the IEE802.11 Committee on wireless local area network (LAN) standards, and in particular is handling standardization of radio communication technology for the personal area network (PAN). Originally IEEE802.15 had planned to develop its own "Mac Lite" system offering simplified IEEE802.11 media access control (MAC) processing, but now the Bluetooth lower layers will probably be used as-is.
There are currently three workgroups in IEEE802.15: IEEE802.15.1, handling standardization of the physical and data link layers, IEEE802.15.2 covering interference with the IEEE802.11 standard, and IEEE802.15.3 working to boost the peak data transfer rate to 20 Mbits/s or higher. Of these, 15.1 and 15.2 plan to complete the standardization process in around March 2001. Once those are done, Bluetooth will hold official status as an Institute of Electrical and Electronics Engineers (IEEE) standard.
At present, 15.3 standardization is planned to be done at the same time as Bluetooth Version 3.0. Tom Siep, IEEE802.15 editor, said, "We are considering using the 5GHz waveband instead of 2.4GHz in 15.3." The Bluetooth SIG has not announced the schedule for Version 3.0 yet.
As standardization efforts begin to bear fruit, the component manufacturers are shifting into high gear: particularly in development of Bluetooth transceiver modules. Most firms are restricting shipments in 2000 to samples, but are probably going to shift over to a volume-production stance from late 2000 through the first half of 2001 ( Fig 4).
Ericsson of Sweden, for example, plans to ship modules with multipoint support by the 4th quarter of 2000 ( Fig 5). By mid-2001, it will sample-ship new modules offering a 40% reduction in mounting footprint, accomplished by various improvements such as integrating the flash electrically erasable programmable read-only memory (EEPROM) in the baseband processing IC.
Of the manufacturers now developing modules, Taiyo Yuden Co Ltd and Murata Manufacturing Co Ltd, both of Japan, seem to be closest to volume shipment. The Taiyo Yuden module mounts everything from the baseband processor to the RF transceiver, while the Murata design holds the RF circuits only.
Taiyo Yuden is ready to volume-ship to meet customer orders from November 2000, commenting,"You have to have a knack for using the Silicon Wave chipset. And we've applied our proprietary waveform-shaping technology to achieve excellent results." Design is said to be underway on a round-the-clock basis, with about 100 design engineers broken down into 10 teams.
Murata Manufacturing will ship RF circuit modules to two markets. The first is for Ericsson transceiver modules, and the second is mobile phone manufacturers other than Ericsson. Manufacturing has started for both types, at tens of thousands of units per month, and plans call for the production volume to be boosted tenfold by the end of 2000.
Moving to Multipoint Connection
Bluetooth-compliant equipment will almost certainly begin with 1:1 (peer-to-peer) connections, for two fundamental reasons.
The first is that the initial Bluetooth applications can be used fully on peer-to-peer connections.
These initial applications will probably be of three major types: connecting mobile phones to headsets; connecting notebook personal computers (PC) and access points (an access point would be a piece of equipment equipped with both subscriber line or Ethernet interface, and Bluetooth wireless interface); and connecting notebook PC and mobile information gear (like personal digital assistants, or PDA).
A source at Matsushita Communication Industrial Co Ltd of Japan said, "Hands-free telephony in the car is very interesting. We expect it to be the Bluetooth application with the highest demand." In Europe many users connect earphones to mobile phones for conversations, and headset manufacturers are hoping to eliminate the headset cable.
Usage in Dial-up Connection
The second type is intended to use Bluetooth in homes and small businesses for dial-up connections from notebook PCs. Already firms like Toshiba, Motorola Inc of the US (parent firm of Digianswer of Denmark) and Widcomm Inc of the US are planning to productize access points for connection to subscriber lines, Integrated Services Digital Network (ISDN) and Ethernet.
Connection between notebook PC and mobile information equipment will concentrate on data synchronization applications. Users with mobile information equipment equipped with Bluetooth functions will be able to connect to their PC via a wireless interface for quick and easy update of scheduler and other data.
Mobile information equipment with Bluetooth functions will probably begin with the expansion module for the Visor, from Hand Spring Inc of the US, and similar systems. Widcomm has already announced it will ship an expansion module for the Visor in the second half of 2000. The firm is also looking at wireless connection for its access point and USB interface adapter (USB Dongle), to ship at the same time. Xircom Inc of the US is also planning to ship an expansion module for the Visor.
Another reason why peer-to-peer will be the most common configuration to start is that the Ericsson transceiver module now being considered by most manufacturers only supports 1:1 connection. This is because the baseband processor does not support multipoint connection, although a new chip which does support it is under development at Philips Electronics NV of the Netherlands: the PCD87750 (development codename Blueberry). Philips has said it plans to sample-ship this month, but it is unclear if this is a realistic target date or not. And if IC shipment is delayed, then it is likely that shipments of new Ericsson transceiver modules supporting multipoint connect will also be delayed.
New Mobile Phones Applications
Bluetooth-compliant equipment offering peer-to-peer connection will support a gradually-increasing range of systems, and eventually end up in most mobile phones as well.
As more and more mobile phones support Bluetooth, more applications will appear to make use of it. For example, using Bluetooth to interconnect home telephones and mobile phones, it makes possible to use mobile phones as cordless phones.
According to a survey by the Communications Industry Association of Japan, even in homes equipped with wired telephones, some 48% of outbound calls and 68% of inbound calls are via mobile phones. And if Bluetooth makes it possible to use the less-expensive wired phone rates with the convenience of a mobile phone, it would mean a considerable advantage for Bluetooth-compliant mobile phone models.
Mobile telephones in Japan are not expected to offer applications for handsfree conversation or dial-up until the second half of 2001. Mobile phone service operators such as NTT DoCoMo and the J-Phone Group are apparently getting ready to add Bluetooth-compliant mobile phones to their line-ups from mid-2001.
DoCoMo has said it will offer a Bluetooth model in spring 2001, at the same time as it launches its International Mobile Telecommunications (IMT) -2000 service. It seems possible that the release date will be pushed back, however, by delays in Bluetooth chipset development.
Access Co Ltd of Japan, which is developing browser software for use on DoCoMo's i-MODE mobile phones, is also developing "NetFront/CompactNetFront with Bluetooth," a browser package for mobile phones equipped with a Bluetooth protocol stack. As a result, it is possible that Bluetooth may be supported by i-MODE phones before IMT-2000 has geared up.
KDDI of Japan and the J-Phone Group, perhaps stimulated by DoCoMo's activity, are also showing a burst of energy. KDDI plans to productize mobile phones using the MSM3300 code division multiple access (CDMA) chipset manufactured by Qualcomm Inc of the US, in the spring of 2001.
The MSM3300 integrates Bluetooth baseband processing functions, which means the manufacturer can make a Bluetooth-compliant mobile phone by just adding the RF circuit module.
The J-Phone Group plans to release a mobile phone with a Java execution environment in mid-2001, intending to use J-Blend from Aplix Corp. J-Blend is scheduled to use the Bluetooth protocol stack developed by Open Interface, Inc. J-Phone intends to use J-Blend in both IMT-2000 and the existing personal digital cellular (PDC) phone networks, so that Bluetooth will be accessible from both.
Personal Network Access
Once peer-to-peer connection has achieved widespread use, primarily through mobile phones, the next phase will be increasing use of applications using multipoint connection.
Toshiba is aiming at small-scale local area network (LAN) applications linking up multiple notebook PC. At meetings, for example, individual notebook PC could form a LAN via Bluetooth to share files and meeting materials. Other firms, including Microsoft Corp and 3Com Corp, both of the US, share their views.
Bluetooth itself is most interested in multipoint connection, and applications which improve connectability: users with Bluetooth-compliant equipment should be able to swap data easily within a range of several meters. 3Com refers to this region as the Personal Connected Bubble, sometimes also referred to as the Personal Information Bubble. Just by walking along the street, the user would be able to pick up publicly-accessible information from passers-by, and make personal information available. This is one application of the personal area network (PAN), and already development has started on games and new business models to utilize the new potential. Users would not need to be aware of contact between their bubbles and others, and could simply achieve data exchange. This capability cannot be achieved, of course, until Bluetooth-compliant equipment is in widespread use, which will take some time.
by Hiroki Yomogita
Alps Electric: http://www.alps.co.jp
Cambridge Silicon Radio: http://www.cambridgesiliconradio.com
Communications Industry Association of Japan: http://www.ciaj.or.jp
Conexant Systems: http://www.conexant.com
Hand Spring: http://www.handspring.com
Infineon Technologies: http://www.infineon.com
Lucent Technologies: http://www.lucent.com
Matsushita Communication Industrial: http://www.mci.panasonic.co.jp
Matsushita Electronic Components: http://www.maco.panasonic.co.jp
Matsushita-Kotobuki Electronics Industries: http://www.mke.panasonic.co.jp
Mitel Semiconductor: http://www.mitelsemi.com
Mitsumi Electric: http://www.mitsumi.co.jp
Murata Manufacturing: http://www.iijnet.or.jp/murata/index.html
NTT DoCoMo: http://www.nttdocomo.com
Open Interface: http://www.openinterface.com
Philips Electronics: http://www.philips.com
Philips Semiconductors: http://www-us2.semiconductors.philips.com/index.html
Silicon Wave: http://www.siliconwave.com
Texas Instruments: http://www.ti.com
(October 2000 Issue, Nikkei Electronics Asia)