@tweet_tree performance report

Posted on Saturday, December 26th, 2009 in web platforms by Ian

@tweet_tree, our Twitter controlled Christmas tree, will be online and accepting your tweets for the rest of today. A performance and traffic report follow.


The tree went online on Friday, December 18th. There were problems reconnecting to Twitter after the first few tweets were fetched. We looked through the code and found a few places where a null-length string might cause problems with the TCP/IP stack. We added a check for null-strings, and reduced the speed from 80MHz to 40MHz as an added precaution.

The @tweet_tree hardware has run continuously since the last firmware update on December 18th (now 8 days) with any errors. That’s about 35,000 search queries (one every 20 seconds) and thousands of tweets processed. By using Twitter as the user interface, the tiny embedded web platform even survived the onslaught of a Slashdoting.


USTREAM transmitted a live video stream of the tree from a web browser/Flash app running on a laptop. The Flash/browser-based system is pretty slick. We noticed that our webcam picks up red, purple, and blue the best, while yellow, orange, and green don’t look as great on the live stream.

We started the USTREAM broadcast on Monday, December 21, and had to reset a frozen laptop about once every 48hours. Unfortunately it froze on Christmas night (25th), and was offline until we reset it the next morning.

The latest USTREAM traffic report is shown above. There were about 11,000 unique visitors over five days, with nearly 105 days of broadcast time. Is it appropriate to broadcast a three-foot plastic Christmas tree to the world, consuming the equivalent of 105 person-days?

The broadcast will continue until the end of December 26th. Thanks to everyone that made our holidays brighter with an ever-changing light show!

This entry was posted on Saturday, December 26th, 2009 at 8:56 am and is filed under web platforms. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.

Leave a Reply

Notify me of followup comments via e-mail. You can also subscribe without commenting.

Recent Comments

  • Drone: So these go for around $330 USD and NO published calibration procedure online or otherwise? Given what's inside - I don't think I'll be considering...
  • Pekka Akselin: This is ridiculous!? :-) We are back at 256(!) byte EPROMs that needed multiple, a handful, of voltages to run! :-(
  • KH: Let's try a back-of-envelope calc balancing energies. From MCP1700 datasheet, there are graphs for a 200mA load step. Estimate the energy shortfall as 12uJ. Say...
  • Daniel: It's been a week and my comment is still awaiting moderation. Apparently the CIA doesn't want their involvement known?
  • KH: Agree, so okay, I guess he must have learned from somewhere. 100nF and 1000uF is so far apart, that was jarring; it's more magic incantation...