I couldn’t help but notice the irony in yesterday’s events. At our Structure 09 Conference, just as panelists started discussing the challenges of managing web infrastructure, we heard that TMZ.com and Twitter had buckled under the massive traffic load that resulted from the news of pop star Michael Jackson’s cardiac arrest and eventual death. (Related: NewTeeVee has put together a collection of his videos online.)
Many went on to proclaim that the web had failed. It seemed a tad sensationalist to suggest that, for only a handful of sites went on the blink. I didn’t hear that Google, Facebook, Yahoo, Microsoft or MySpace broke down. For them, it was business as usual. Nevertheless, this morning, I decided to check two sources that monitor the performance and traffic patterns of the web. From web monitoring service Keynote Systems:
Beginning during the late afternoon pacific time major news sites, including those of ABC, AOL, LA Times and CBS all started to show marked slowdowns in performance. Beginning at 6:00 p.m. Eastern time until 8:00 p.m. Eastern, the following sites showed became nearly unavailable (down to nearly 10% availability): ABC, CBS, LATimes, AOL, CNN Money.
Beginning at 5:30pm (EDT), the average speed for downloading news sites doubled from less than four seconds to almost nine seconds. During the same period, the average availability of sites on the index dropped from almost 100% to 86%. The index returned to normal by 9:15pm (EDT).
I got further details from Akamai, which operates a global content delivery network that gives it a unique view into traffic patterns on the Net. Around 6:30 p.m. EST, Akamai’s Net Usage Index for News spiked all the way to 4,247,971 global visitors per minute vs. normal traffic of 2,000,000, a 112 percent gain. Of this, 3,556,495 visitors were accessing sites from within the U.S., Akamai said.
Akamai’s data shows that the traffic started to build around 5:30 p.m. EST and peaked just before 8 p.m. The big spike to online news sites appears to have lasted about 20-30 minutes. This morning, the traffic to these news outlets remains higher than the daily average; at last check it was around 3 million visitors worldwide.
The data collected by the two agencies shows that sure, there was massive spike in the traffic that lasted for a little while. And then things pretty much returned to normal.
Thanks for the sanity Om. No break downs here at WordPress.com either.
Yeah, you are right about that Budd. I think it is a testament to your team’s good work. WP, it just works.
Yes it does. And I love what WP has done with the latest releases.
It returned to normal once people got over the shock, shrugged and went “Oh well, MJ was pretty much old news by now anyway”.
Well Google did break. This morning I was trying to run a web search on something completely unrelated and got a Google error basically indicated they thought they were under DOS attack, and it wouldn’t let me search without entering a capthca which didn’t work on my crappy Nokia E51 phone.
Om – This post is right to put in context yesterday’s failures, but it’s sloppy in two places. First, it’s not ironic that this occurred during your infrastructure discussion. Perhaps it’s serendipitous. Also – The timescale uses EDT, EST and times without a noted offset This makes it hard to follow.
While i was doing some work on Facebook.
I checked in on Twitter.
Everyone posted tweets about MJ’s death.
I felt shocked and really felt sad.
But something went wrong with my Twitter profile yesterday.
Now it’s ok.