Reflection Attacks on Server

Question

I was reading a research paper entitled "Steps to Defend Against DoS Attacks" when I came across this point:

"A reflection attack on a server would require server-client-server communication, and most clients are not going to respond to a new connection request from a server and hence reflection attacks on server will be prevented"

It was written under the heading of Application of IPv6 addressing aiming to prevent DoS attacks.

My question is that how server-client-server communication is needed for reflection attack on a server?

I Googled for it but couldn't get to answer.

Answer 1

A 'Reflection-based DoS Attack' in the traditional sense means that a number of computers spoof their source IP address to the same address (10.1.1.99) who is supposed to be the intended victim, and then the computers connect to a 'reflector' (10.1.1.88). The reflector attempts to respond to all those connection attempts by connecting to the spoofed address, but ends up flooding the victim with unsolicited traffic.

With IPv6, the addressing scheme is designed to prevent spoofing because a portion of the address is supplied by the network(s) the machine is a part of, so these types of attacks could not be launched.