ComputerWorld notes that 802.11g will have 10 to 20 Mbps throughput, not 54 Mbps, but...: This has been known all along. The 54 Mbps rate is the raw symbol rate, or the number of symbols per second including all framing, error correction, etc. The net throughput rate of 10 to 20 Mbps (or higher) erases all the stuff you need to carry data and looks at the payload.
The IEEE has already talked about how, in future standards, they might be able to bring the throughput up closer to the raw symbol rate. This story is pretty close to on target, but should have described the differences between throughput and raw data. Manufacturers have been saying up to five times faster, and I've been saying: yeah, if you get 4 Mbps out of 802.11b (a real-world figure for a network with multiple users) and 20 Mbps out of 802.11g (an idealized figure for a non-mixed b/g network with a single user).
This is true of Ethernet in general anyway. On a gigabit switched network you can only expect 400-600Mbps throughput after all the overhead.
I'd like to see you post some commentary on this. I've been reading some of the talk on slashdot about this and i think some folks are confused as to what this really means.
slashdot article:
http://slashdot.org/article.pl?sid=03/05/22/2216247&mode=nested&tid=137&tid=193
thanks
It's quite possible to get 980Mbps on Gbps Ethernet, since the signaling rate is 1.25Gbps.
Well... we've certainly never seen 980Mbps and we have our network layout optimized for speed and bottleneck issues. Speeds are certainly significant faster when sending only a few very large files vs. a huge amount of tiny files. We're running all CAT6 cabling throughout our entire studio in heavily shielded conduit.