The Case for Strong Authentication of Network Traffic

The openness of networks and the Internet has undoubtedly led to the success and growth of public networks. New applications and features have flourished out of the lack of strict security requirements and the anonymity offered by public networks. However, this openness has also been arguably the single greatest enabler of annoyance and malicious use of networks and the Internet. Spam, denial-of-service attacks, address spoofing, routing attacks, and a myriad of other malicious uses are, at least partially, the result of allowing unauthenticated network traffic. This leaves network applications and protocols on their own to implementing security and attack prevention. In this essay, I will argue that we need to take at a different approach: strong authentication for network traffic. Network traffic authentication will enable security and protection from applications that demand it while still allowing for innovation in network applications.

Strong authentication of network traffic would provide a means for all traffic to be verified for integrity and identity of sender. Currently, most network traffic, especially IP traffic, provides no means for verifying the sender. IP specifically only requires a source address to as means to identify a sender. This is far from a strong form of authentication as IP addresses are easily faked. A strong authentication system, on the other hand, may utilize public key cryptography and multi-factor identification or some new technology to guarantee the identity of a sender and integrity of data sent. It is not my intention to argue in favor of a particular network traffic authentication mechanism but rather to show that one is needed and indeed feasible.

Problems with Unauthenticated Traffic

The lack of an authentication mechanism has created the opportunity for malicious use of networks, particularly the Internet. The ability to uniquely identify a sender is fundamental to preventing unwanted and malicious network traffic. Currently, the Internet relies on source address as the only form of identification of an IP host. However, through IP spoofing and routing hijacking, individual hosts may change their identity by changing their IP source address. IP spoofing is the forgery of ones source address in a packet. Routing hijacking attacks are much more sophisticated attacks in which malicious users broadcast illegitimate routing information in order to take control of IP address space.

This identity agility makes it difficult to prevent many of the types of attacks that are common on the Internet. Prevention systems often make use of blacklists, lists of IP addresses of known malicious users, to block unwanted traffic. Spam mitigation is a great example of using blacklisted addresses to identify known spammers. However, the effectiveness of blacklisting is hampered by the fact that host can change their identity. Strongly authenticated network traffic solves this problem by replacing the ambiguous IP address with an identification mechanism that uniquely identifies an individual or organization in way unchangeable by a sender.

Clearly, this is only one example of where the lack of authentication creates problems on the Internet. IP addresses create a safe haven for malicious users to hide behind while performing an inexhaustible list of undesirable network actions. This total anonymity also has social implications which may be solved by strong authentication of network traffic. Although, this is not to say that an authentication system would eliminate all anonymity and privacy. It is easy to imagine a system in which a trusted third-party verifies your identity on your behalf and would only give out personal information to authorized individuals.

The Push for Authentication of Network Traffic

One common argument against authentication of network traffic is the stifling of innovation and creativity on the Internet. One of the reasons new and innovative uses have emerged for the Internet is the openness and ability to reach any host by default. Opponents can argue that adding strong authentication to network traffic will slow innovation because of the removal of anonymity and universal reachability. However, this is not necessarily the case. Although network traffic is authenticated, this does not mean that there is trust between a receiver and sender. Applications and protocols that do not need a strict trust relationship can simply authenticate all senders and continue functioning as before when network traffic was not authenticated. Applications that benefit from trust relationships will use strong authentication to their advantage. A clear example of this would be a public website authenticating any client that wants access regardless of whether or not the site trusts that client. On the other hand, BGP routers could take full advantage of strong authentication to ensure that they are communicating with trusted peers, preventing routing attacks.

Authentication has already made its way into many network uses paving the way for universal