This is a page to describe what features we want in a (layer 2) wireless protocol. 802.11 technology has proved very popular but has limitations. Assuming that 802.11b will be replaced with some other protocol at some point, it makes sense to be proactive and start laying out some requirements for this protocol. Please add any requirements or features to this page, and if it becomes a comprehensive set, it can become a working outline for a real charter for a new protocol that we can put before hardware suppliers. So far, this idea is in the concept generation stage, and it makes sense to mine the problems we have with 802.11b to do concept generation for a new protocol. When the new protocol inevitably does emerge, this will give us something to gauge it against.
- Improved range over 802.11b
- Mesh features
- Ability to service many clients without losing bandwidth for each user
- Ability to operate efficiently in ad-hoc (AP-less) modes
- Bandwidth shaping
- Wireless Vlans
You may or may not know that 802.11b has already been effectively replaced, by 802.11g, which features improved throughput via a modulation scheme called OFDM, and that many of the other problems you describe are not solvable on the layer 2 level. Continued research is being done, and new protocols, even in the 802.11 family, are being finalized. We've just gotten a glut of new technology and the industry needs a while to let them soak.
Bandwidth per channel (as opposed to per client) is a characteristic of all RF communication - I believe Vivato's switch devices are an attempt at solving this. Mesh is a concept which is already being investigated in earnest by anyone with any interest. Traffic shaping and virtual networking is already being effectively implemented with 802.11b and other current protocols, by any router which is capable of such things. Efficient ad-hoc operation is another fundamental problem with RF; it suffers from another often-discussed problem known as the hidden node, which causes networks to degrade quickly with significant numbers of nodes.
I think anyone who is capable of seriously approaching these problems is already working on them, although your enthusiasm for new technologies is welcome. I'm sure any specific ideas you might have would be very welcome to the groups who are working on implementing them.