News

Bluetooth shock delay to version 1.2 -"hopefully this year"

It was seen as a rubber stamp: when CSR announced that it had silicon conforming to Bluetooth 1.2 standard earlier this year, it was assumed that the new products would have these important features by this Christmas. Wrong! - the spec may not even be published by then.

The warning given in eWEEK two months ago, was confirmed with an official announcement in the Bluetooth monthly magazine, Incisor, which attempted to put a positive gloss on the news. "As many of our readers will know, the 27th of September was the due date for the Bluetooth Special Interest Group (BSIG) to publish its update to the core specification – moving the Bluetooth world on from version 1.1 to 1.2" says the magazine in the latest edition.

However, on the 1st of October the Bluetooth SIG gave the bad news: "Interoperability testing for this version of the specification is now complete and ... the testing results also show that Anonymity mode, Absent mask, and Scatter mode are not mature enough to be included in the next draft of the Bluetooth Core Specification."

The good news is good: "The results of these tests confirm that Adaptive Frequency Hopping, eSCO, Enhanced Inquiry Scan, L2CAP Flow and Error Control and revised QoS are mature features."

But the bottom line is, version 1.2 isn't approved, and won't be for a while. And there is bound to be dissent in the ranks, because for many SIG members, the features which are not yet ready, are not vital ones.

The full online edition of the magazine includes an interview with Anders Edlund, marketing director of the Bluetooth SIG, in which he explains the importance of the delay.

Edlund is quoted: "Anonymity mode allows you to mask a device's address, meaning that people – or their Bluetooth products, anyway, can't 'see' a device. This is about privacy and security. The absence mask is about the way a Bluetooth device defines the period of time when a device will talk to another. For example, devices on a heavily-loaded network can be disconnected for a period of time to reduce loading, and then re-connected. The work we are doing on scatter mode concerns the way several small Bluetooth networks send information between themselves. This update will improve/enhance the way that this works."

For the sake of these relatively minor features (said one disappointed source) Bluetooth will now not be able to handle quite vital problems, like adaptive frequency hopping.

Edlund's interview concedes that AFH is important, because it relates to co-existence of Bluetooth with WiFi: "More and more computing devices – whether they are notebooks, PDAs or other devices such as Tablet PCs – now feature Bluetooth and WiFi in the same box. Now, Bluetooth suffers less from interference in this scenario than fixed frequency wireless standards such as Wi-Fi, but we felt it was right to take the initiative and add AFH to the Bluetooth spec. This will benefit other wireless standards, too," he said.

Dissidents say that if Edlund and the SIG understand the importance of AFH, why was it thought worth delaying this upgrade for the sake of getting scatter mode and absence mask and anonymity?

There will, almost certainly, be some members who will advocate trying to rush the version out early. Stalwarts of the SIG, however, say it won't be as easy as it was, for example, for Broadcom to jump the gun in the WiFi race, when the silicon maker released 802.11g hardware onto the market nearly a year before the WiFi Alliance could do comparisons, and before the IEEE could even ratify the standard.

"The difference between us and WiFi is that we all share the IP," said Nick Hunn of TDK Grey Cell. "Nobody owns the design, and if you tried to get out early, you'd be breaking patents owned by the SIG. We all agreed to share all the intellectual property when the SIG was set up."

One good reason to believe that pre-release silicon will be a small problem, is the fact that the most attractive market Bluetooth vendors can see, is a very long-term one - either factories, or motor-cars.

Car makers are looking at building Bluetooth standard products into vehicles which will ship in three years' time, said SIG president, Mike McCamon, at a recent meeting of the SIG in London. Waiting a couple of months, now, doesn't affect them at all; but launching a standard which might cause them problems in 2006 would be a disaster. The same applies to factory equipment, currently being designed.

ESCO is about providing improved voice quality, especially in noisy environments. The headset companies and mobile phone were pushing for this, and, as we know, headsets are a very important Bluetooth market opportunity. ESCO tackles this with new package types.

Enhanced inquiry scan will provide faster connection and set up times. As it becomes more commonplace to be operating a Bluetooth-equipped device in an environment where there are many others, this becomes increasingly important.

"As far as L2CAP flow and error control is concerned, this will help people that are sending large files around. L2CAP helps avoid errors in transmission – even one wrong bit can throw a printer out, for example," Edlund says in Incisor. "And finally, we pushed to complete the revised QoS. Improved QoS is becoming increasingly important as people are mixing more and more Bluetooth devices together – on the desktop, for example, when you can be using a Bluetooth keyboard, mouse and a voice connection all at the same time. In this scenario you need fast connections and good voice quality."

Edlund refused to be drawn on when the 1.2 spec would be finalised.

Asked if it would be out before the end of 2003, he said: "At this point we do not have a date that we will publish for inclusion of these features, and so it is possible that this could create some issues or at least delays for developers. But – and this is very important – I will remind you that we are staying resolutely with the goals of the '5 minute ready' programme. It is all about the quality of the out of box experience and we will not compromise the situation by including new features in the spec that are not sufficiently tested."