|My Take on the Future of Wireless Sensor Networks||| Print ||
|Written by Akiba|
|Wednesday, 11 November 2009|
I’m taking a break from my seemingly neverending hardware assembly process because I was suddenly inspired to write a blog post about the future of WSNs as I see it. I’m not an analyst, nor would I want anyone basing their decisions off of something I said. However I do have the luxury of meeting with quite a few people in the wireless sensor network community (both Zigbee and non), hanging out with a rather large crowd of uber geeks (Tokyo Hackerspace), and am exposed to WSN news just about every day of my waking life for the past two years from updating the blog’s news feed. So in a way, that makes me slightly more qualified than some dumbass research groups (like ABI Research) that pay their interns to surf the web and collect datasheets, parts specs, and articles to put together a multi-thousand dollar research report based on publicly available data.
The past two years that I’ve been heavily involved with wireless sensor networks have been very interesting and, I daresay, exciting. I’ve actually been involved with WSNs on a very superficial basis for longer, but the last two years got me kind of deep into the middle of the pile. When I first started collecting news, it was to try and aggregate the disparate and sparse information that would occasionally surface. Today, I find myself spending a lot of time every day sifting through all the wireless sensor press releases, blog posts, and related information that seems to be flooding the internet. The verdict: wireless sensor networks’ popularity is increasing at a frighteningly fast pace.
When I first started, most of the wireless sensor news was coming out of research or academic institutions, and pretty much only Zigbee or other vendor-specific solutions were available. I first heard about 6LoWPAN a little while after I started my open source Zigbee project although it existed for a while before that. At the time, there was a bit of back and forth protocol bashing between the two camps and I was also taking part in it. It reminds me of the whole east coast/west coast conflict in the US (at least in the hip hop community) back when I was a dancer. You would take a side, just based on where you lived without really understanding what the fight was about.
Anyways, two years spent deep in the WSN industry, combing through all the WSN news, and watching the industry grow exponentially ages you fairly quickly. A lot of my previous views on different protocols have moderated as I’ve watched some rise, some fall, and became weary of protocol wars. Being exposed to groups that are tech savvy but outside the WSN industry like Tokyo Hackerspace has given me a lot of different perspectives on the technology as well.
I’m not just rambling because of nostalgia though. I believe that the WSN industry is getting close to a big change where it will be converging with IPv6. A lot of people in the Zigbee Alliance are saying that Zigbee/IP will just be for smart energy, but I don’t think so. WSNs are growing to the point where there is a lot of corporate interest in them now, as well as consumer interest. You regularly hear about Google, Microsoft, Cisco, IBM, Intel, and many other companies talking about wireless sensor networking which would have been huge news just a year or two ago. My exposure to Tokyo Hackerspace has also taught me that there is a huge wealth of experience and knowledge about IP networks, and coincidentally, they’re interested in wireless sensor networks but don't know where to start. There are also new wireless sensor protocols coming out of the woodwork.
The only obvious solution would be to converge WSNs at the network layer to IPv6. This would be the magic bullet that everyone is looking for. Corporate interests would go wild over this because there is already a huge infrastructure for IP networking that not only spans the internet, but reaches into people’s homes. It would also make things much easier for consumers to adopt because it’s a familiar protocol. Although every house might not have an IPv6 expert, there will be a lot of information on the internet for people to look up problems within a short time after something like this would be introduced. And finally, it would put an end to the eternal protocol wars and bashing between all the different camps. It’d be really difficult to argue with an IP solution. Some would probably discuss a protocol's energy efficiency, but you just can’t argue about infrastructure. An extra three hours of battery life doesn’t compare to instant access to billions of consumers.
There are still issues with a transition to IPv6 and I suspect next year will be the year that a lot of them will be ironed out. App layer issues need to be worked out such as discovery mechanisms. Routing needs to be decided on, and I believe there is still debate on both sides about which type of routing to use. Security and encryption for resource constrained devices will also be required. There needs to be top layer protocols for device to device communications and interoperability. And finally, test houses and test specs need to come online to verify interoperation in order to make things painless for the consumer.
Interoperability is one of the toughest issues and its one area that Zigbee does well in. Although a lot of people complain about Zigbee interoperability, its probably the most mature and furthest along with many test specs, well defined application profiles, generic application libraries, test houses, and plug fests. For now, you can pretty much assume Zigbee/IP is a feeler for how well IPv6 (6LoWPAN) meshes with the Zigbee top layers. But you can pretty much bet that once the issues are worked out for the energy industry, all the other profiles will rapidly follow. In my eyes, IPv6 for wireless sensor networks will be the future, and WSNs will be a physical interface for the internet.
|< Prev||Next >|