It was a great day to be at the beach in San Francisco (Apr 27, 2008). Thanks to Jason Stone and Jim Stockford for sparing their Sunday afternoon, we met up at Ocean Beach with three XO laptops. The little green machines attract a good bit of attention. Even before we got started with out little experiment, we had people stopping by to ask questions or just check out the green machine.

[1]

[2]

 

 

So, the experiment was: How far will the point-to-point mesh network on the XOs go in relatively clear RF (radio frequency) environment? At least cleaner as far as 2.4 GHz spectrum goes. We did such a run on campus and got about 600 feet max. or so across two XO units. I suspected that this was because our campus has a lot of Wi-Fi traffic, so the interference alone would decrease the hop length. We also have a lot of foliage, so that would add to the interference.

Ocean beach is right along the Pacific at the edge of San Francisco. I hoped that RF would be cleaner there. So, we got started by setting up the mesh of three laptops (me, Jim and Jason) and then got Jim firmly planted (say, position A) in the thick of beach population to answer questions. Jason and I walked south while pinging Jim's laptop all along. Well, we kept on walking...and walking...and walking. This was definitely way beyond 600 feet. Eventually, we started seeing "host unreachable" a few times, so we stopped. We saw ping times of about 25 to 30 ms. Jason became point B there. I kept walking further south and after a few steps, the pinging resumed with low latencies (3 to 5 ms). I kept walking...and walking...and walking. A quick note about walking on the beach. Walk closer to the water. Its a lot firmer and works well when measuring paces (1 pace approximately equals three feet). Eventually, I started seeing "host unreachable". I called Jim to verify. He was seeing the same. I tried holding the laptop higher up and the pings would resume. Eventually, I stopped at what I thought was the outer limit of pings, marked the spot and took the following picture.

[3]

 

Walking back from my point (C) I counted 649 paces to point B. I walked along the water's edge to keep my paces approximately equal. So, that's 1947 feet. Over three times what we saw at SF State campus. Next, we walked from Jason's post (point B) back to Jim's point A. That was 575 paces or 1725 feet. Almost three times the observation at SF State. I noticed that at Jim's post, we were still close to houses and condos along the beach and saw a lot more APs on the neighborhood view.

[4]

 

At point B I saw one AP and at point C I saw no APs at all. So, the interference at point A must have been a lot more than at B and C, explaining better range between B and C.

That was triumphant. Ironically, the wireless feature of the XO is one of two non-FOSS bits of code in the laptop. There is some movement [5] to replace it with a FOSS firmware, but we'll have to wait and see.