Performance Test: 802.11b Takes a Lickin' and Keeps on Tickin'

by Rob Flickenger

With all of the talk about the performance impacts of WEP (Wired Equivalent Privacy) and noisy environments, I thought I'd throw some quick, informal benchmarks together to see if any of this hearsay was true.

To test this, I generated a 1,441,792-byte file (as 1,441,792 bytes x 8 bits per byte = 11,534,336 bits, or 11 megabits) by copying it from /dev/urandom using dd. I then tried sending this file to an IBSS (ad hoc, also known as peer to peer) node like this:

$ time dd if=random.bits |ssh -c none  dd of=/dev/null

This essentially means "Take these 11 million random bits, copy them over the network using no encryption, just dump them into the bit bucket on the other end, and tell me how long it took you." I did this to try to minimize the impact of disk usage and CPU crunching, and just try to make the bits fly as fast as possible.

I tried sending the file at 1, 2, 5.5, and 11 megabits per second, from a couple of different positions: about 25 feet away from the node (through three walls and a non-operational microwave oven), the same with the oven boiling some tea water, and then from about 20 feet away in the same room as the operating microwave oven. I did this with WEP encryption off, on, and at Lucent's 128-bit RC4, all without external antennas.

Comment on this articleHave you ever experienced microwave oven interference, or any other type for that matter, with your 802.11 network?
Post your thoughts here

The first interesting thing that I noticed was that, no matter how hard I tried, I couldn't squeeze out more than about 4 Mbits. This could be due to a number of factors, the most likely being that I'm using a cheap PCMCIA/ISA bus adapter in my router (a Pentium 233), and the bus probably just couldn't keep up. I'm really not so worried, as my Internet connection is only 1.544 Mbps max anyway! ;)

Will microwave ovens blow your 802.11 network?

Will microwave ovens blow your 802.11 network? Our tests say no, but your mileage may vary.

Before we get to the numbers, I'd like to point out that even with the above fancy command, there was still a small amount of system overhead in actually getting the packets sent. As the exact amount is difficult to calculate but non-trivial, I decided to weigh the figures like this:

At 1 Mbps, it should take 11.00 seconds to transmit my 11 Mbit file, in the best possible case. On average (each test was sampled five times and averaged), it took 14.91 seconds to complete.

So, we have 3.91 seconds of unaccounted-for overhead (or 35 percent of the total transmission time.) For purposes of argument, we'll assume that the 1 Mbit speed is optimal, and deduct 35 percent from all transmission speeds (chalking it up to system overhead.) And so we are grading on a curve.

At any rate, here's the performance at 25 feet, through three walls and a solid wooden door, no antennas, behind the unplugged microwave, without WEP:

1 Mbps: 14.91 - 3.91 = 11.00 seconds = 1,048,576 bps (adjusted)
2 Mbps: 8.33 - 2.91 = 5.42 seconds = 2,128,106 bps
5.5 Mbps: 4.82 - 1.68 = 3.14 seconds = 3,673,355 bps
11 Mbps: 4.26 - 1.49 = 2.77 seconds = 4,164,020 bps

The same with 40-bit WEP:

1 Mbps: 15.09 - 5.28 = 9.81 seconds = 1,175,773 bps
2 Mbps: 8.36 - 2.92 = 5.44 seconds = 2,120,282 bps
5.5 Mbps: 4.89 - 1.71 = 3.18 seconds = 3,627,149 bps
11 Mbps: 4.60 - 1.61 = 2.99 seconds = 3,857,637 bps

And again with 128-bit RC4:

1 Mbps: 15.05 - 5.26 = 9.79 seconds = 1,178,175 bps
2 Mbps: 8.37 - 2.92 = 5.45 seconds = 2,116,391 bps
5.5 Mbps: 4.86 - 1.70 = 3.16 seconds = 3,650,106 bps
11 Mbps: 4.66 - 1.63 = 3.03 seconds = 3,806,711 bps

As you can see, WEP appears to have virtually no effect on throughput, until cranking 11 Mbps. The difference in transmission times was just about 1 percent. I believe it can be chalked up to natural variances (not to mention user error!) At 1 Mbps, WEP performance actually tested slightly better!

There was a bit of an impact at 11 Mbps (about 8 percent or so) when going to 128-bit encoding. Not sure what that means, as it wasn't even getting anywhere near 11 Mbps without WEP to begin with. Really, the highest figures should probably just be tossed until I figure out why it isn't transmitting efficiently.

As the difference in throughput was negligible, I left 128-bit encryption on for the remaining tests.

Next came the fun part. 300-watt microwave oven on at full blast, nuking a cup of water, while trying to upload an 11 Mbit file:

1 Mbps: 27.64 - 9.67 = 17.97 seconds = 641,866 bps
2 Mbps: 13.30 - 4.65 = 8.65 seconds = 1,333,449 bps
5.5 Mbps: 7.05 - 2.46 = 4.59 seconds = 2,512,927 bps
11 Mbps: 6.02 - 2.10 = 3.92 seconds = 2,942,432 bps

At 1 Mbps, I saw the worst performance when directly behind the running microwave oven, at a 45 percent performance hit. But at 11 Mbps, in the same setting, the performance was only down about 23 percent.

Related Articles

Recipe for a Linux 802.11b Home Network

802.11b Tips, Tricks, and Facts

An Introduction to Lucent's Wavelan Wireless LAN Card

Using SSH Tunneling

Granted, that was a highly contrived test setting. I think the ambient room test (about 20 feet from the node, with a microwave oven chugging along in the same room) was more representative of what you can expect in the "real world":

1 Mbps: 16.86 - 5.90 = 10.96 seconds = 1,052,402 bps
2 Mbps: 9.15 - 3.20 = 5.95 seconds = 1,938,543 bps
5.5 Mbps: 8.52 - 2.98 = 5.54 seconds = 2,082,010 bps
11 Mbps: 5.07 - 1.77 = 3.30 seconds = 3,495,253 bps

There we have it: 11 percent loss at 1 Mbit; 8 percent loss at 11 Mbps.

So, in the absolute worst case, inches behind a running microwave and at 1 Mbps, you're still pulling 640 Kbps -- about double the speed of the average DSL line, and about 12 times the bandwidth of a "very fast" dial-up connection!

I can't wait to test the above against Bluetooth equipment running in the same environment -- assuming it ever makes it to market. ;)

Incidentally, the tea was delicious (although a bit too hot.) Mmmm ... Yerba Buena Maté.

Rob Flickenger is a long time supporter of FreeNetworks and DIY networking. Rob is the author of three O'Reilly books: Building Wireless Community Networks, Linux Server Hacks, and Wireless Hacks.

Return to the Wireless DevCenter.