Timothy D. Morgan’s recent paper titled, “IPv6 Address Cookies”, seeks to apply the fundamental shift in resource availability brought about by the vastly increased Internet address space in IPv6 to develop a novel, lower cost solution to mitigating spoofed attacks. “Spoofed denial of service attacks have plagued the Internet for a number of years, and show no signs of abating. Research into mitigation techniques has apparently not led to a financially viable solution, and new attacks have been discovered in the wild without being widely anticipated”. The following provides an introduction to this paper:

It has long been known to researchers that address spoofing on the Internet is a serious problem. While a great deal of effort has been put into finding theoretical and practical solutions, spoofed attacks are still globally endemic. They represent a simple nuisance to many, but a business-halting bane to others.

Enter IPv6. IPv6 is the next generation of the Internet protocol designed to alleviate the existing global address shortage and improve the scalability and extensibility of the aging IPv4 protocol. This new protocol provides a huge 128-bit address space which should provide enough addresses for a great deal of Internet expansion. Individual Internet users can easily obtain their own 80-bit block of addresses if they currently have a single IPv4 address. This means it will be possible for any user to effectively hide a system in such an address space without ever being found. That is, if no DNS records point to them, and they don’t respond to broadcast requests, it would be effectively impossible to find a host through brute force probes.

To put this into perspective, let us suppose an attacker can scan addresses of a network at 2³² (more than four billion) packets/second (which is probably a stretch for today’s fastest routers). It would take on average, 2⁴⁷ seconds (or 4.5 million years) to find a single live address in a range of 2⁸⁰ addresses. For this reason, the only way to find systems that do not pick an obvious IPv6 address would be to query some name resolution service (such as the DNS). In this paper, we propose methods which utilize the large IPv6 address space to mitigate spoofed attacks by forcing clients to always use the DNS prior to submitting service requests.

Further information on this research and the paper produced is available on the SCN website.