Priority Interrupt by Steve Ciarcia Do the Math

I have a Panasonic pan/tilt/zoom (PTZ) web cam at home so that when I’m not there, I can watch UPS leave packages on my front deck and check out 2' of snow in the driveway and wonder if my plow guy has taken a hike. In fact, the web monitoring works so well that I decided to install a few web cams at my vacation cottage so I can watch things there too.

The web cam at home runs on hard-wired Ethernet. I have high-speed DSL at the cottage; however, I mostly use Wi-Fi there because there’s no easy way to string wires. So, slap in a wireless-G router, plug in a couple indoor/outdoor Toshiba PTZ wireless (802.11b) web cams, and I should be golden, right? Can you spell “frustration”?

I installed the first wireless web cam and it seemed to work fine on the intranet. It worked so well that I even decided to set it to 1,280 × 960 resolution (the IK-WB11A also has 800 × 600, 640 × 480, 320 × 240, and 160 × 120 settings). My Panasonic camera at home has a fixed 640 × 480 resolution. Seeing the higher resolution was encouraging. Smiling with success, I spent the rest of the afternoon mounting a camera high up on the front of the cottage under the overhang. The only wire to it was a simple DC power cord. It worked like a charm.

 If one works well, then two will work better, right? I mounted a second camera on the back of the cottage and set it for the same resolution. When I turned on the system though, one camera ran at about one-tenth its previous speed and the other came up once and stalled. The cameras were about 50' from the Wi-Fi. Hmm. Because I was working with 802.11b, I thought perhaps the range was somehow limiting the already limited bandwidth and it couldn’t handle two cameras at 1,280 × 960. I lowered the camera resolution to 640 × 480 and everything seemed to perk up. OK. I could live with that. The picture update wasn’t really fast, but it was reasonable.

 Next, I wanted to see how well this would all play when I looked at things from my house or the office. I went to a local Wi-Fi hot spot to see how the images were piped over the Internet. Well, I could get a slow picture on one camera, but the other one just came up with the display page and froze. I could understand the massive file size of 1,280 × 960 (about 1.2 MB raw) choking the pipe, but 640 × 480 (about one-quarter the file size) should have been a breeze. Bandwidth tests showed that my DSL was running at 2,144 kbps for downloads and 318 kbps for uploads. Something was jamming the system. I figured it must have been a bandwidth problem because of that stupid “B” rated wireless. Where’s “N” when I need it?

 OK. I threw up my hands and decided to hard-wire the front camera directly to the closest Ethernet switch in order to eliminate the wireless bandwidth jam. Curiously, I got pictures from both cameras, but the updates were still too slow to be useable. It took stringing an Ethernet cable room-to-room directly to the router to discover that the current Ethernet switch was too slow. The second camera also still needed a better Wi-Fi signal (no way to hard-wire this one).

Finally, things started looking better after I installed a high-gain antenna kit ($69), five-port Ethernet gigabit switch ($79), and a Linksys Wireless-G broadband router with SRX ($179). At the hot spot, I could at least get both cameras to update slowly, but they didn’t work at the same time. Disappointing, but useable. Then it hit me. Do the math, stupid.

We’ve gotten so used to hearing advertising hype about high-speed Wi-Fi and streaming video that it’s easy to forget that they are talking about your intranet. Outside your high-bandwidth multimedia intranet is the same old DSL Internet connection that is pretty fast one way and excruciatingly slow in the other direction. I may be able to download data at 2,144 kbps, but the best I can do for downloading images from my web cams at the hot spot is 318 kbps—the DSL upload speed of the web cams. A 640 × 480 picture is 640 × 480 × 8, or about 2,457 Kb. At 318-kbps upload speed, that’s about 7.5 s per picture! With 50% compression that’s still at about 4 s, and I haven’t discussed processing delays, protocol overhead, and slow links. I won’t even think about 1280 × 960 resolution anymore. I’m lucky I get any pictures at all.

Seriously though, it finally ended up working enough to be useable, but just doing the math wasn’t necessarily an accurate predictor. For the same resolution setting, the Panasonic motion JPEG at my house seems to update at about 0.8 s per frame. The Toshiba java applet pictures from the cottage refresh at about 2 s per frame. Both DSL services give the same results in bandwidth tests. At this point, I’m not going to look a gift horse in the mouth. Whether by enchantment or skill, at least it works. However, if any bright person with real expertise in this area would care to shed some more light on this subject, I for one would like to be enlightened.