So, just looking for a sanity check here. I’m getting ready to deploy a new network, and I’ve been very happy with results in the lab! I am just curious to hear feedback and suggestions from the community.
The center of the network is a 200 foot tower with 9dBi horizontal polarized 2.4Ghz omni, and 12dBi vertical polarized 5.8Ghz.
The 2.4Ghz network is a standard B/G subscriber AP, which should be just fine.
The 5.8Ghz network is my “feed” network, where multiple sites will connect to for data backhaul. Before you comment on it, I cannot do point to point for each tower because this is a co-location tower and I can only have 2 antennas.
The actual data feed into the 5.8Ghz network is a client on the AP itself, not ethernet into the AP. I do realize that the hub AP in the middle will have to do double work on all packets, but this is something I’ve taken in to the network design. The network is currently configured with WPA2, Nstream/polling, and 5Ghz-turbo.
In all testing I have reduced TX-powers and introduced loss to acheive signals in the -50 to -60 range, SNR is ~50dB noise floor is -99.
I am able to transmit about 30Mbps through the network end-to-end from a p3 700Mhz windows laptop running btest to a p4 2.2Ghz windows pc. If I am considering everything properly, that means I am getting 60Mbps from my primary AP. All of this is without maxing out any single CPU in the middle. Total throughput results are the same for unidirectional or bidirectional tests.
It seems to me that this is “pretty damn good”, and I should be so lucky as to see even half of that out in the field.
Does that sound like a fair assessment to you all? Any questions, suggestions, or comments are welcome.