Today I got my first ever taste of the Slashdot Effect from the other side. A fluffy little rambling I wrote voicing some concerns about the Silk web browsing component of Amazon’s Fire tablet actually got linked by my old home base. I didn’t submit the story myself- it was submitted by someone completely at random. And then the traffic started.
Had I known I was gonna get Slashdotted, I probably would have bothered proof-reading my story too. I just tossed it out of my head unedited, not really intending it for serious publication. One of the very first posts on Slashdot noticed a grammar error in the lead graph. Oops. Hah. Some things never change. The comments were fun. I miss that place more than I could ever express.
But what of the Slashdot Effect?
Over the years I’ve seen many estimates of the true scale of The Slashdot Effect. The initial link could trigger from 5k to 100k page views depending on the interest of the story. But beyond that, the secondary effect of the link can be far larger: a curious reader might click around and load a couple more pages beyond the actual link. And perhaps more important, they might pass that link along to their friends in broader networks beyond Slashdot’s normal reach. Of course, I’ve never been able to measure anything for myself.
There’s a lot of variance of course: clicks depend on how interesting the story is. My story was about security/privacy implications of an unreleased tablet computer. It scores some points because it’s Android, Amazon, and Kindle, but I wouldn’t call it a home run story. Some percentage of Slashdot readers would probably RTFA, and the fact that it was written by me might have drawn a few extra friends in just to see what I’m up to a month after resigning from Slashdot.
So anyway, in the 5-6 hours following the appearance of my link on Slashdot, the original link got about 5k page views, and the rest of the site got about that many more. The Slashdot story itself got about 113 comments which is probably a little below average, and while my link in the article was the lead link, there were actually 6 links within this particular story to distract people, which undoubtedly drives the total clicks down to each individual link, so I’m guessing my numbers are a little south of typical for a small scale Slashdotting in 2011.
But for me it’s pretty neat. It’s also worth noting that my VM handled the load nicely, weathering 10s of thousands of HTTP requests without breaking a sweat, spiking up to about 80% CPU utilization briefly, but then falling back down to around 40% for the subsequent few hours. Not bad for what it is, which is a cloud hosted box.
I still wouldn’t consider myself ready for a more substantial traffic explosion, but I think my sense of scale on what defines “Serious” is a bit different than most. To me, if I’m not dealing with a thousand pages a minute, it just doesn’t seem like a big deal… a few thousand pages an hour just seems so trivial. Then again, I’m hosting it on amazon’s EC2 using their free micro VM. The price sure is right.