Earlier this week, YouTube suffered an outage which lasted for
almost two hours, a serious amount of time for the second most popular
website on the Net.
The whole Internet went crazy with rumours and theories, and
the news spread like wildfire, popping up on Twitter
almost immediately. After the Amazon S3 outage earlier this month, I’m
sure
everyone was wondering if YouTube was to blame or
Google themselves
had gone down.
What Happened?
We now know what happened;
Pakistan attempted to block the video-sharing site due to what it
considered to be offensive material about Islam.
In doing so, they
hijacked some of the IP addresses directing traffic to YouTube,
propagated the hijacking to other DNS servers, and essentially routed a
good chunk of YouTube’s audience nowhere.
This Could Happen On A Larger Scale
What’s mind boggling about the incident, and BusinessWeek/ZDNet
Asia
picked up on, is that there is nothing in place to prevent this from
happening again, or shut it down quickly when it occurs.
DNS routing is
based on an elaborate trust system orchestrated by ICANN, who states
that they lack the ability to pull an offending server out of the
equation.
In other words, the entire Internet is a elaborate
interweaving of
trust and any given location can turn at any time, and we have nothing
more than a toothless lion in ICANN managing a master list of which
server has which AS number.
Conclusions
Most webmasters pay more attention to
server security than any governing body appears to be paying to the
entire virtual infrastructure of the Internet.
This was the result of what is allegedly a simple mistake; a
quick-fix change to ban a site for an entire country that was then
accidentally propagated to a partner that then created the outage. But
just imagine if it had been intentional and directed at more than one
site.
This article is based on Profy post written by Cyndy Aleo-Carreira.
