Teething Pains or Cavities?
Hiawatha Bray wrote about his experiences with the first set of Bluetooth devices. Reality highlights the Bluetooth's limitations (full essay).
12-Nov-2002Version 2: 2023-01-31 17:29:21

Hiawatha Bray wrote about realities of Bluetooth in the Boston Globe today (Nov 11, 2002). He did a fine job of summarizing the realities of Bluetooth though I may be biased since it confirms what I've been saying.

Note that you can read a summary of these comments on SATN.

Bluetooth has indeed succeeded in its goal of being a wire-replacement. It has reproduced all the problems of wires just without the wires. The tangled mess might be invisible but you still need to worry about pair wise matches between all the devices and you are still limited by the length of the imaginary wire. In addition Bluetooth creates limitations of its own on how many nonwires can be connected to an access point (7, you don't even need to use your toes to count them) and on the amount of data each nonwire can carry. It's very strange that one must not use a wire in order to get the benefits of Bluetooth. The nonwire is a very short leash whereas the Internet, wired or not lets us connect any two points anywhere in the world, not just within 10 meters.

We should learn from the example of X.400. X.400 was (is?) a mail protocol approved and required by essentially all the telecommunication agencies throughout the world. It was designed over a period of ten years yet failed against SMTP (Simple Mail Transport Protocol) which could be implemented in an afternoon. Like x.400, the Bluetooth was designed and promulgated before anyone could learn from the first generation. Bluetooth is designed to work in the specific cases imagined by its designers and thus will perform very well in precisely those scenarios and these are the scenarios touted in press releases. It's not surprising that if you don't use Bluetooth precisely as envisioned it will not work very well. There is a tendency to view these problems as anomalies and those of us who point them out are considered spoilers and are thus discounted.

If we are clever enough to work around the problems of Bluetooth then Bluetooth is viewed as an enabler rather than an impediment. For example, Microsoft is only using Bluetooth as a transport for Internet connections, albeit one that is slower than 802.11. This appears to be an endorsement of Bluetooth rather than being a triumph over it that allows all those Bluetooth radios to be repurposed to support the Internet. For the sake of this discussion I'm considering the use of Bluetooth purely as an IP transport to be essentially 802.11 with a different radio. It's similar to use a modem to transport data despite the voice phone network.

Bluetooth itself is not very interesting and would not be worth writing about except that it diverts effort from creating more effective and simpler solutions. Bluetooth is essentially a version of IrDA (the red plastic infrared port on most laptops).

The problem with IrDA (and Bluetooth) is that it tried to give us solutions rather than the tools to create our own solutions. If you wanted to hold your laptop in front of an IrDA printer you could indeed print a page. But if you wanted to simply share files on a network you could not do that. The tragedy is that we could have had wireless connections nearly ten years ago. Now we have IrDA redux in the guise of Bluetooth and we have the same problems.

Not only have we been entangled by wires for these past ten years, we've lost valuable time in addressing the problems raised by wireless connectivity. We would have seen the naive futility of firewalls which simply melt away when anyone can connect with a corporate network. The pair-wise security model of Bluetooth simply doesn't work in any but the most trivial scenarios.

We must have encrypted IPv6 as a basic capability in order to make untethered connectivity simple enough to work.

Deploying 802.11 raises the same issues yet I'm disappointed at the degree to which these issues are ignored. Instead users are told to manage complex and arcane security regiments which serve to fragment the commons rather than create effective communities.

Though Bluetooth and 802.11 are very different most people don't understand the underlying technology for either one and they tolerate the problems with 802.11 just like they accept that Bluetooth will one day deliver on its promises. To most the technologies appear unrelated. Bluetooth is about eliminating wires and 802.11 is about home networking. (To add confusion, 802.11 is also marketed as WiFi).

My challenge is in explaining how they are different and why it matters. It's the same challenge I face in explaining why telecommunications policy matters.

Bluetooth is in the mainstream of the old model of telecommunications in which all the services are defined by the center and every new capability must be approved before it can be deployed and thus before we even understand it. 802.11 is simply a transport for packets and doesn't stand in the way of creating new capabilities.

Once again we face a familiar paradox. Bluetooth which defines so much of the solution is thus limited to what it defines and that is very little and it only works among a few nearby devices. 802.11 which makes few promises inherits the existing richness of the Internet Protocols and has no such limits of distance.

The problems Hiawatha is finding in experimenting with Bluetooth are basic to the design and not exceptions. The flaws are inherent in shipping what amount to applications rather than extensible technologies. Basic assumptions such as the need to share cell phone radios are problematic. In fact, for my laptop, I simply have a second cellular radio on a PCMCIA card and thus can talk on the phone while exchanging data. Unlike land phones where I pay for service independent of phones, with cellular service one pays for each radio. It's like when motors were very expensive so a factor would have one motor and drive all other machines using a complex set of belts and pulleys.

Bluetooth's security model, replication model and other applications have their own flaws.

But we should not spend any time trying to solve any of these problems. Microsoft's approach of using Bluetooth as an IPv6 transport is the right one and lets us take advantage of the entire Internet -- both the net itself and what we've learned.

More important, it allows us to start to address the real issues we face as we move beyond the static wired Internet to the untethered Internet. While 802.11 gives us a taste of what is possible we must recover the Missing Internet with encrypted IpV6 and dotDNS.

The blue e's in the headline are copied from the style used in the Boston Globe's summary page. The headline writer seems to have discovered that one can be cute online as well as in print.