Tackling Internet routing security

Most Internet users have experienced or at least heard about visible security threats such as spam or phishing attacks.

These days, they are also pretty well equipped with anti-malware defences that give some level of protection and security against such attacks.

However, there are other threats, which are linked to the core functioning of the Internet. These threats are less visible to the end-user, are potentially critical, and almost impossible for client-side computers to detect and defend against. These are security breaches which can affect Internet routing security.

The majority of network relationships are based on a system of mutual trust. Users trust that the route used to transmit the information will not be altered in any way. This trust model, however, is open to attack by various means, the most common involving scenarios where false information is injected into the routing system. This can result in potential outcomes such as black-holing, impersonation, inspection and alteration, and network destabilisation.

For example, in February 2008, end users accessing YouTube were unable to reach the site because an inadvertent routing leak initiated in Asia caused a complete global disruption of service for two hours. Errors such as this would have been detected if the route chosen had been between 'trusted' Internet resources.

Geoff Huston, chief scientist, APNIC, says, "The Internet has become a key business platform and now has so many transactions occurring online that these security risks have real cost implications for businesses."

The most robust response to this issue is through Resource Certification. To ensure that false information is not injected into the Internet's routing system, parties attest or certify that the routes they advertise to transmit this information are genuine and that they hold a valid IP address and who is resource registration.