Want to hear a fun story?<p>You remember dyndns.com ? Well, it used to be a small app run literally from a dorm room at WPI. Well, after a while, it became one of the largest dynamic DNS providers in the world, and router manufacturers around the world started supporting it in the UI. Dyndns used to provide an address (an olden-days plaintext web API of sorts) to check the external IP of a router. Eventually, manufacturers realized that even if they didn't have a deal with dyndns or even formally support them, they can use their IP address to check the router WAN IP. Major manufacturers like linksys and d-link (I think, don't take my word for it) jumped on the bandwagon and everyone started pinging the crap out of that page and if they took it down, these routers would break in interesting ways. So they built it up and kept it running.<p>That API runs to this day, at checkip.dyndns.org.<p>edit: Today that company is dyn.com, bootstrapped pretty much all the way, and provides DNS to twitter, etsy, fastly, the guardian etc.
It's very easy to cause a ton of problems with commodity routers. The University of Wisconsin - Madison CS department got DOSed back in 2003 by Netgear routers using the CS NTP server because it was hard-coded into the router firmware. Full story is linked below:<p><a href="http://pages.cs.wisc.edu/~plonka/netgear-sntp/" rel="nofollow">http://pages.cs.wisc.edu/~plonka/netgear-sntp/</a>
Am I correct in thinking they use a single site hosted by them to decide if the router is connected to the internet? Did noone not think this might not be the <i>best</i> idea?
This event reminds me of an old definition for "distributed system":<p>"A distributed system is one in which the failure of a computer you didn't even know existed can render your own computer unusable." (source: <a href="http://research.microsoft.com/en-US/um/people/Lamport/pubs/distributed-system.txt" rel="nofollow">http://research.microsoft.com/en-US/um/people/Lamport/pubs/d...</a>)<p>In this case, the failure of a system most people didn't even know existed (Belkin's heartbeat server) rendered their router unusable.
My guess? Belkin was routing DNS queries from these boxes through their own infrastructure. For what purpose, who knows.<p>Evidence is the "workaround" posted on their status page:<p>"We have identified a workaround that will enable some users to get back online. The workaround requires that you set a static DNS address on the device trying to access the internet."