Duke Research Experiment Disrupts Internet Traffic

alphadogg writes with this excerpt from Network World about an experiment gone wrong which affected a big chunk of internet traffic yesterday morning: "It was kicked off when RIPE NCC (Reseaux IP Europeens Network Coordination Centre) and Duke ran an experiment that involved the Border Gateway Protocol (BGP) — used by routers to know where to send their traffic on the Internet. RIPE started announcing BGP routes that were configured a little differently from normal because they used an experimental data format. RIPE's data was soon passed from router to router on the Internet, and within minutes it became clear that this was causing problems. ... [f]or a brief period Friday morning, about 1 percent of all the Internet's traffic was affected by the snafu, as routers could not properly process the BGP routes they were being sent."

Re:A big chunk?

[dreamchaser]

Ok, so what if 1% of all people on the planet just dropped dead? That would be over 60 million people.

Now apply that ratio to the thousands of nodes and terabytes of data that flow over the 'net. One percent is quite a bit more than you seem to think it is.

No I'm not comparing it to people dying. I am just illustrating the point.

Re:A big chunk?

[Voltageaav]

If they had, you probably wouldn't even know about it because you wouldn't be reading slashdot. You'd probably just be calling your ISP asking them why your internet connection isn't working.

Re:A big chunk?

[linebackn]

1% can be either large or small depending on what is being measured. For example, a Pointy Haired Boss may think the following:

* 1% of your web site user base using a different web browser is insignificant and can easily be ignored.
* 1% of your annual profits is HUGE and losing or failing to obtain those means heads must roll.

(of course, a true PHB will never see any potential relationship between the two)

Re:A big chunk?

[thePowerOfGrayskull]

That's kind of the point isn't it? 1% isn't; a few hundred million is. That's the risk of using percentages: they tend to minimize the significance of the real numbers -- or alternatively, overstate their significance.

Re:Not surprising

[phyrexianshaw.ca]

the trusting nature of BGP? if you inject bad advertisement for too long, you'll get marked as damaged, and the BGP AS will begin converging around you. (cutting off any private subscribers you maintain as they no longer have valid routes back to them from the internet.)

in current BGP, you don't GET trusted, you BUILD trust. you're established a very high metric (or weight) for distance routing initially, and as you carry traffic, (or as more and more traffic originates from your network from your subscribers) your metric will be lowered overtime, moving greater and greater volumes of traffic over your infrastructure.

Terrible article, bad summary

[Eunuchswear]

What a lot of verbiage to say:

Some routers have bad BGP implementations that handle attributes longer than 255 bytes incorrectly

Some of those routers will drop a BGP connection if thet get such an attribute.

The article makes it sound as if RIPE is in the business of distributing routes to BGP routers.

(See http://www.merit.edu/mail.archives/nanog/msg11505.html [merit.edu] for details).