Connecting Devices With Wireless Grids

Roland Piquepaille writes "A new concept is emerging in networking: wireless grids. These grids connect all kinds of wireless devices, such as sensors or cell phones, with each other and with more traditional wired grids. IEEE Internet Computing has devoted a very long and thorough article about these wireless grids which can deliver new resources, locations of use, and institutional ownership and control patterns for grid computing via ad hoc distributed resource sharing." (Read more below.)

"The article says that applications for wireless grids fall into three classes: the ones which aggregate information from the range of input/output interfaces found in nomadic devices, those which focus on the locations and contexts in which the devices exist, and those that leverage the mesh network capabilities of collections of nomadic devices. The authors add that these grids "emerged from a combination of the proliferation of new spectrum market business models, innovative technologies deployed in diverse wireless networks, and three related computing paradigms: grid computing, P2P computing, and Web services." If you're interested in the future of wireless networks, the original article is a must-read, but check this summary if your time is limited."

Comments

Printable version...

[tcopeland]

...is right here.

Re:connecting...

[igrp]

I was going to hold off and mod in this thread but there's one point I would like to make instead.

You address a valid point and your attitude is probably shared by many. However, there's also the bigger picture which few people take the time to look at (and I don't mean any single person).

These wireless grid concepts (at least the more sophisticated ones) are basically scalable, distributed computing solutions. They solve a lot of problems but also suffer from some of the problems that all distributed computer networks have.

The more common wireless grid devices become (provided this idea ever takes off) we'll likely see two major changes: on the one hand, efficiency will increase dramatically (more devices = a bigger ressource pool, common protocols, less overhead). And on the other hand, we'll see a change in how we view CE and mobile computing in general. Thing about it: most of the time our PDAs, cell phones, etc. don't actually do anything, but rather just idle.

Those spare CPU cycles could, however, be used by others in the grid which would in turn require their device to be less powerful (since they can depend on the network's CPU power and need to do less computing onboard).

There are three potential big problems I see with this though:

• Battery life. The more our devices actually 'work', the more power they consume. Add the network/radio overhead which generally tends to consume quite a bit of power on top of that and you have a problem. People have to come to accept a certain standard of usability though. So you either need to limit power consumption, and hence cripple a mobile device's effectivity as a grid note or add bigger batteries.
• Manufacturers. They basically won't like this. It takes a less powerful device to accomplish the same task. So why buy a new gizzmo if you can just use some nearby grid node.
• And lastly, sheer economics. For this to work, people would have to see benifits now. Nobody wants to buy a new PDA that has less battery life just to be part of something big if there's no immediate gain. And, due to the fact that you need a whole bunch of grid nodes in order to have a useful network of some sort, that just won't happen.

  "Hmm, if I buy this now I might get free Internet access two years down the road. In the meantime, everybody's freeloading off of me though (since there's nobody else whose device you could use). I think not."