Feeds
|
Location:
Mon, 03 May 2004
The Hype of "Mesh Networks"
I have to admit, the concept is pretty cool. I throw 30 "devices" in random Seattle locations, each discovering one another; creating relationships of some sort; and being able to transmit information (seamlessly, of course) from one end of the cloud to the other. If a few die, it would route round them. If there was a building in the way, it would find another path to send my information. And, last but very not least, this mesh network would provide such high QoS, I could send voice and video over it. And then we fall back to the real world -- none of this idealistic crap exists. Ouch, that hurt. Sure, there have been many attempts.....but no one has come to the table with a true, innovative, open solution. And no, configuring a bunch of radios to random IP addresses (just hoping we avoid guessing the same number) doesn't count. With the state of technology today, we may be asking for too much. First, we are burdened with half duplex radio technology, ie: 802.11. Strike 1. In order to have a true, dynamic mesh, we need omni directional antennas. Strike 2. Efficient distributed trust systems, resource allocation, QoS, and lack of spectrum to provide this. Strike 3. You're out! Its the omni-mesh paradox, and until we can somehow configure multiple full duplex pairs while talking to multiple hosts on the fly, I dont see our issues going away. MIT RoofNet (and a few other solutions popping up on the scene) might mitigate some of these hurdles. But if there was a widely available mesh routing solution, trust me, we'd be running it. I do have to say that some routing protocols, like AODV, are some nice attempts at mesh routing -- even though they have little to do with wireless in particular. I am not a big fan of stateful on-demand routing, but its a great concept, and has its place in the world. For past blog entries, check out the archive on the side or click here. |
Make some extra cash with your blog too: |
