Judging by my e-mail, many of you are convinced that Microsoft has done nothing but steal and cheat its way to success. While some in Redmond might admit to stealing an idea or two, there are a lot of Microsoft folks who would take exception to the "nothing but" part. For example, the Simple Object Access Protocol (SOAP), which will undoubtedly play a significant role in the Internet's future, is an idea hatched in Microsoft's labs. That's pretty innovative, if you ask me.
Based on my discussions with Microsoft wireless architect Mike Foley, who also chairs the industry consortium that oversees Bluetooth (Bluetooth SIG, I'll go out on a limb and say that Microsoft is showing signs of innovation in the wireless space as well.
Yesterday, Microsoft released to manufacturers an OS update that will bake support for Bluetooth right into Windows XP. When combined with WiFi (802.11a/b-based wireless Ethernet), Microsoft will call this its "Wireless Desktop"-- a PC that literally has no cables save a power cord. Microsoft believes the wireless PC will spawn new innovations from systems vendors in terms of the PC's design and form factor.
But it's not Microsoft's reaction time to a burgeoning technology that's innovative. What's innovative is Foley's vision for where wireless technologies like Bluetooth fit into the grander scheme of things. For example, Foley already seems determined to turn Bluetooth's device-specific profiles of today (what Bob Frankston called baggage in my previous column on this topic) into a legacy that only requires ongoing, perhaps dwindling support tomorrow.
Not only does Foley understand where Bluetooth best fits into the world, but he also understands why that future should be inextricably linked to version 6 of the Internet Protocol (IPv6) and what sort of innovations can result from such a marriage. Prior to chairing Bluetooth SIG, Foley chaired one of its working groups. The focus of that working group was Bluetooth's Personal Area Network (PAN) profile, a part of which is IPv6 support. "Microsoft will support legacy profiles to support existing devices," explained Foley, "But our feeling is that all new innovation for Bluetooth should be done using PAN and IPv6. This way, Bluetooth automatically inherits any of the benefits that come from improvements to the Internet's protocols, and solutions that work with them. For example, if I make a printer, I would like to use the same software to connect that printer to the network regardless of what physical transports are used."
To make that vision a reality, however, Foley's PAN working group had to get under the hood of Bluetooth to overcome some of its weaknesses. Bluetooth is a master-slave technology, and that aspect of it is still there in the PAN profile. It's also a TDMA-based technology where the slaves (usually devices like keyboards, mice, etc.) can only "speak" during timeslots allocated to them. The master (usually a PC or an access point) is in charge of that allocation. But, with IPv6 riding on top of a master-slave system like Bluetooth, it's up to the master to make sure that none of the slaves are starved for communication. "To that extent." said Foley, "I think we mapped IP onto it reasonably well."
But, according to Foley, one thing that's not in the Bluetooth specification that you might find in other wireless technologies like WiFi is the ability to natively handle multicasts, where one node can talk to many nodes with a single transmission. "There's a workaround to the problem in the current PAN profile," said Foley, "but Bluetooth takes a performance hit for using it. If we're on an Ethernet segment with four PC's and one PC does a multicast, the rest hear it and process it in the same slice of time. With Bluetooth's master-slave architecture, it serializes a multicast into a series of unicasts that each require their own slice of time."
Issues like this led me to recommend last week that you start looking for IPv6 support in your Bluetooth devices. In general, a slave's ability to handle the PAN profile is closely tied to the vendor's choice of radio. Foley suspects that most slave-like devices (keyboards, mice, cell phones, PDAs, etc.) can't be updated with a software upgrade the way some master devices like PCs and access points can. But even with the latter, it's not a guarantee. You have to check with the vendor. During a recent briefing with a vendor of mobile devices, I asked the product manager if the built-in Bluetooth devices supported the PAN profile and he said "I think so." But I could tell by the look on his face that he didn't even know what I was talking about. Getting this information from your vendor may be tricky. (My future reviews of these devices will be sure to unearth this exotic but critical detail.)
At the most recent Bluetooth World Congress in Amsterdam, Foley wowed his audience with a four-node Bluetooth network consisting of a mouse, a keyboard, a PC that also had WiFi connection to the Internet, and a DVD player. The DVD player was connected to the PC via IPv6-enabled Bluetooth. Therefore the DVD player was a full-feldged node on the IP network with a few smarts (like an HTTP server built into it). To prove it, Foley entered "MyDVD" as the URL into a browser running on the XP machine and the result was a display of the contents of the CD in the DVD player. On its own, the DVD player had retrieved the CD's songlist off of one of the many CD lookup services on the Internet and piped it back to the PC via HTTP over the Bluetooth connection. Clicking on one of the songs in the browser with the Bluetooth mouse brought the DVD player to life and, from there, Foley said, "Look ma, no wires" and played back a bunch of songs.
What's the big deal, you ask? I guess it doesn't take much to dream up this scenario. But someone had to do it first and that someone was Microsoft. And Foley, who is both a Bluetooth activist and Microsoft employee, was rightfully in the thick of it. If you ask me, that counts as innovation.
I asked Foley about a wireless technology lurking on the horizon called UltraWideBand (UWB) that many claim will snuff the lights out of WiFi and Bluetooth. Proponents of UWB say it's not only faster than anything out there, it's also cheaper and consumes less power. Said Foley: "UWB is just another reason that support for IP is so important. If the applications running on Bluetooth, WiFi, and UWB all support IPv6, then upgrading the wireless layer won't break those applications." I must admit that hearing "standards" talk like this coming from Microsoft is a bit unusual, and it's another reason that I think the company may, for a change, deserve an "A" for innovation.
When I asked Foley if he saw himself as an innovator, I was once again surprised. Instead of the brash, overconfident response that I'm used to hearing from Microsoft personnel, he said, "Let's just say I've spent a lot of sleepless nights thinking about this."
Note to Steve Ballmer: Keep Mike Foley awake. Get him a lifetime membership to Starbucks.
Do you agree with David? Is Microsoft innovating? Or has someone pulled the wool over is eyes? TalkBack below or write to david.berlind@cnet.com.
