Saturday, May 31. 2008802.11n - The official delivery of pipe dreamsTrackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
Don't forget the whole mess of the CSIRO patent thing on parts of 802.11n..
Hmm, your comment system stripped the link, lets try it again as just text..
http://arstechnica.com/news.ars/post/20070924-dark-australian-patent-cloud-looms-over-802-11n-spec.html
Good pickup Chris. I had indeed forgotten about the CSIRO patent claims. (I'm guessing that the IEEE would also like to do just the same!)
It is indeed a mess, and something that does need to get sorted one way or another.
The CSIRO claims apply to 802.1a and g too, which have receive diversity modes which use the CSIRO technology. 802.11n just more blatantly infringes, and thus is the test case the lawyers are taking forward.
No one will be able to demonstrate 110Mbps. Remember that wireless is half-duplex and TCP Acks every second packet. So the link needs to be turned around too frequently to get 110Mbps. Your HD streaming comment is unwise. Video can be sent in one direction only, and we don't care about loss (since the codec will cope with a certain amount of that). So HD streaming is one of those niche applications likely to work well (as you might expect, considering that HD video is broadcast on wireless today, via DVB-T). Your ethernet cable will outperform wireless. TCP's performance is very sensitive to loss and cable has none of that, whereas loss is inherent to wireless. The headline bandwidth is less important than the loss rate, and it's a shame the article you quote didn't measure loss as well as bandwidth (although the decrease in connection speed further from the access point is an attempt to limit the loss rate). MIMO also promises less worst-case performance. It's difficult to think of an objective test to measure that. But a baseline comparison against a good 802.11g access point would have been interesting. I would be careful about the number of points installed in your house. On one hand, installing as many ports as possible during construction is cheap and a good idea. On the other hand, you're looking at about $500 per 24 ports to light up a domestic outlet at GbE and having unlit ports on the wall is a pain in the butt. |
QuicksearchCalendar
Blog StatsLast entry: 2008-08-19 21:37
211 entries written
225 comments have been made
Blog AdministrationBlog Tags alerting apps birthday building coding cricket entertainment family football friends gnome google hardware health holiday house Internet ISP life linksys Linux marriage microsoft monitoring movies nagios networking news office pets play review security shopping sms snoring spam sport support tennis terrorism web weekend windows wine work |
|||||||||||||||||||||||||||||||||||||||||||||||||