Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I need to be able to connect a computer to our network with realistic speeds to the rest of the network of 30+ Mbps (as measured by something like iperf). Running a wire is not an option here, and the computer is about 50 feet away from the rest of the network, through a floor and wall.

There are no other sources of WiFi or any other 2.4Ghz or 5Ghz interference in the area.

I've thought about trying 802.11n / WiFi N to do this, but I cannot find realistic benchmarks of its performance. Most benchmarks are at very short distances. I'm looking for a benchmark where they actually test the performance at different ranges to see how the bandwidth drops.

I understand that the 300 Mbps claims of the manufacturers are not true, but is it unreasonable to expect 30+ Mbps consistently?

share|improve this question
1  
I'm sure if you call someone like Cisco and show interest in one of their enterprise grade APs, they'll be more than happy to loan you one for a short period so that you can test. When I was speccing APs for our dorms, I was loaned multiple models of APs. –  MDMarra Oct 6 '11 at 16:01

2 Answers 2

up vote 1 down vote accepted

Very loaded question. Only way to respond is an answer with a mouthful of qualifiers.

From a throughput perspective, 802.11n in a radio and antenna configuration on both AP and client that is capable of 300,000,000 bits/second PHY interface data rate (n can actually go as high as 600 Mb/s as defined in the spec) -- herein referred to as 300 Mb/s (network interfaces are rated in base 10, decimal using standard K, M, G prefixes unlike file sizes and memory which use base 2, using Ki, Mi, and Gi prefixes) in a green field with single client with line of sight at fifty feet without interference will be able to achieve 30 Mb/s at both the PHY and at the application layer with UDP L4 transport or TCP L4 transport with a modern OS tuned TCP implementation and traditional LAN latency of less than 5ms. Application layer throughput of 30 Mb/s puts PHY throughput > 30 Mb/s. Maximum sustained throughput will likely be higher, but 30 Mb/s sustained is very feasible given the qualifiers.

Reduce RSSI or introduce spectrum interference which ultimately introduces latency which ultimately throttles down TCP and whatever your green field maximum is for a given system begins to drop.

With that said I use Cisco 1242AG access points (a/b/g capable, not n) in both a and g paradigms and am able to sustain between 20 - 25 Mb/s application layer throughput (SMB, CIFS, FTP) on the local LAN (out of 54 Mb/s possible at the radio PHY) with a single client at about fifty feet through a single interior wall. However, spectrum interference does cause noticeable drops in throughput.

Moving from a/g to n will only increase the maximum throughput and reliability. However, it is wireless -- and until you test it -- all bets are off.

share|improve this answer

I understand that the 300 Mbps claims of the manufacturers are not true, but is it unreasonable to expect 30+ Mbps consistently?

How consistent is your "consistently"? I'm gonna assume you mean as consistent as your wired infrastructure.

You need to appreciate that radio waves can very easily be affected by lots of things in the environment and as such its impossible to guarantee anything. If the connection absolutely positively must always be available and must always be running at a certain capacity or else, then wireless is probably not for you.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.