1-year Anniversary of Leo Laporte using TalkShoe

It’s our 1-year anniversary of Leo Laporte using TalkShoe, and time to reminisce…

It was back in October 2006 when Leo Laporte of TWiT-TV first told us that he was planning to use TalkShoe. Leo, for those of you who don’t know, is the #1 podcaster in the world by most measures. We had a couple of weeks to prepare, and we had no idea what kind of traffic Leo would generate, and what part of our system would break. At that time, we didn’t (we do now) for example have anyway to heavily load-test our system.

The BIG day came on November 5, 2006 when Leo and Amber (MacArthur) first went live with their net @ nite (http://www.twit.tv/natn) Talkcast (previously called Inside the Net). The whole company (all 10 of us) came into the office on that Sunday night, as we prepared for the 9:00 pm live Talkcast. We all popped open some beer, and as the time approached, we could see from our traffic reports that TalkShoe’s system was “heating up.” Web traffic jumped starting about 15 minutes before the show, along with telephone traffic.

Finally, Leo came online and the show started. We all were on pins-and-needles as more and more people joined our TalkShoe Live software. Leo and Amber talked about TalkShoe and their show, and the chat room was hopping with multiple chats every second or two. Finally, at about 20 minutes after the hour, the TalkShoe website went down! Leo and Amber had successfully flooded the TalkShoe system.

We had prepared for this event, and quickly put up a system-down message which we labeled “No Shoe for You” taking off on the Seinfeld episode of the Soup Nazi and “No Soup for You.” What the heck, we figured, let’s at least have some fun with it.

After another 10 minutes or so, the chat room of TalkShoe Live began to freeze. There were just too many chats coming in for the system to handle. We hadn’t anticipated this issue. Of course people kept entering chats saying they couldn’t see their chats, and so on. After another 10 minutes, the chat resumed, and then was slow for the rest of the Talkcast.

Luckily, as we had predicted, the voice/phone/conference part of the system performed flawlessly. That’s because it’s based on a telco-grade conferencing system from Compunetix that is installed in companies like Sprint, MCI, and British Telecom, unlike other Asterisk-based systems that can only handle relatively few calls.

When the Talkcast ended, we all breathed a sigh of relief. We had made it, albeit with some battle scars.

When we spoke to Leo, he was great. We explained our problems, and he was actually somewhat jazzed by the fact that he had “broken” TalkShoe as he said. Most importantly, he said he loved the system, and would be patient and work with us.

Over the course of the next 4 weeks, we made many changes to the system including: 1) implemented load testing, 2) tripled the computer server capacity, and 3) re-designed our software to handle additional load on the website, TalkShoe Live and the chat room.

Now, almost exactly 1 year later, our system can easily handle hundreds simultaneous callers on a single Talkcast, and thousands simultaneous streamers, website visitors and chatters. Oh…and Leo and Amber still like us   and use TalkShoe regularly.