Is this a personal project, or do you plan to integrate it into a commercial device?
If it’s just for 1-2 status pages served up to 1-10 users and lag isn’t a whole lot of concern, then any hardware you choose will most likely work.
The main reason I see that most microcontroller based WiFi solutions only use up to 802.11b is simple baud rate. RS-232 operates at a maximum baud rate of 230.4 kbps. I2C maxes out around 3.4Mbps. If the hardware you’re interfacing can only pull a limited throughput anyway, it makes sense to choose the WiFi module that stays as close as possible to the maximum interface speed, otherwise the excess speed is wasted. Sure you can buffer the output of the wireless interface to compensate for burst and lag, but a microcontroller can’t suck data in as fast as 802.11b can push it.
The other reason, as I see it, is cost. 802.11b devices are cheap and plentiful, and when selling components to experimenters low price point is an important consideration. They work for a lot of situations, and even at 11Mbps it can outpace a T2 (~6Mbps).
bdodds:
I don’t plan on serving more than 1 or 2 pages with a JPG image or two, maybe not even any pages and only JPGs.
I think 5 users at one time would be the max and more likely only 1 or 2 at a time.
This is a personal experimentation project for now.
In that case you can choose whatever you want to use. SFE sells an embedded webserver, there’s the arduino solution you mentioned, and there’s always the option to create something yourself. It all depends on what you want to get out of it. There are also multiple servers to choose from, and multiple ways of serving data (http, ftp, tftp, etc). Nice thing about the web standards is that they’re all completely free to access, so if you want the standard for HTTP/1.1 it’s free to obtain. (Unlike a lot of other SPEC documents.)
bdodds:
Thank you very much for the explanation on why microcontroller based WiFi solutions only use 802.11b. Makes perfect sense now.
No problem...
bdodds:
As for cost, I don’t think $80+ is very cheap when a full 802.11N router could be the same price or less.
Remember with consumer devices like wireless routers you're dealing with economy of scale. A router can be expected to sell more than 200k units, so each one is fairly cheap to make. A component like a wireless module may only sell 10k units over it's revision cycle, so it will naturally be more expensive. $80 is still a bit on the pricy side (IMHO), but for a low demand item, I think it's in the right range.
Please feel free to share your project as you work on it. I’d like to see where you go with it.