From the wikipedia article on Denial-of-service attacks:
A denial-of-service attack (also, DoS attack) is an attack on a computer system or network that causes a loss of service to users, typically the loss of network connectivity and services by consuming the bandwidth of the victim network or overloading the computational resources of the victim system.
In particular, one vector of attack to pay attention to is
consumption of computational resources, such as bandwidth, disk space, or CPU time
This kind of attack is successful because of an inherent asymmetry between the attacker and the attacked. The attacker performs very little computation to send the packets, but the server has to accept them and perform some computation to determine if they are valid or bogus. In this way an attacker with the same or less computational power can overwhelm a bigger host.
These same kinds of attacks are repeated in many different kinds of media where the asymmetry is present. For example sending email is practically computation free, but reading and processing it takes time. The result? Spam. With the advent of many competing blogging services starting a blog is also practically free, and the predictable result is splogs.
There are different kinds of defenses, but the one I want you to keep in mind is removing the asymmetry, making the computation requirements on both sides closer in parity. The importance of this will become apparent in a bit.
The reason I bring this up is because not only do denial of service attacks occur in electronic networks, they also occur in the real world. The one I want to talk about is the "social denial-of-service" attack. This can happen, for instance, in the case of an IETF Working Group where all of the work is supposed to be consensus based. In that scenario it is very easy for an attacker to derail progress with a constant stream of objections and wildly divergent proposals.
The IETF, being the IETF, has an RFC for this.
Notably, in a small number of cases, a participant has engaged in what amounts to a "denial-of-service" attack to disrupt the consensus-driven process. Typically, these attacks are made by repeatedly posting messages that are off-topic, inflammatory, or otherwise counter-productive. In contrast, good faith disagreement is a healthy part of the consensus-driven process.
For example, if a working group is unable to reach consensus, this is an acceptable, albeit unfortunate, outcome; however, if that working group fails to achieve consensus because it is being continuously disrupted, then the disruption constitutes an abuse of the consensus-driven process. Interactions of this type are fundamentally different from "the lone voice of dissent" in which a participant expresses a view that is discussed but does not achieve consensus. In other words, individual bad faith should not trump community goodwill.
Let's go back to electronic denial-of-service attacks. They worked because of an inherent asymmetry between the attacker and the attacked. The same is true of the social denial-of-service attack where arguments, responses, rebuttals and more importantly time has to be spent responding to the bad faith objections, which are easily written up and tossed onto the mailing list.
Remember that one way to fight a denial of service attack is to raise the amount of computation required by the attacker. In the case of a Working Group the way to do that is by requiring disruptions to take more time and energy. This is where the call for "camera ready copy in the form of a Pace" comes from in the AtomPub WG. Camera ready copy is much more difficult to write than a one or two line objection tossed into a mailing list. Only if you are willing to put in the work to write up a Pace with reasonable text will it start to take up the time of the WG. Your willingness to put in the time and effort to create camera ready copy will distinguish your proposals and objections from those of an attacker.
So if you are participating in the AtomPub WG don't be put off if you're asked to put up a Pace with "camera ready copy", you are after all helping in the fight against social denial-of-service attacks.
Many thanks to Mark Pilgrim for the many long talks we've had about this subject. In the course of those talks we've come up with two observations, the first by me:
And a corollary by Mark: