The networking world seems to be losing sight that NAT is a crutch of sorts, a way of dealing with the primary problem of a lack of IPv4 addresses. An earlier article in IPJ stated that NAT provides a firewall function. I think NATs and firewalls are mutually exclusive, even if they are found on the same networking device. This is because NATs don’t by themselves provide any natural protection from the host on the other side of a protection point. The two can operate independently.
NAT does present real-world problems with a number of products, such as Microsoft AD Replication and IBM’s Virtual Tape Library. Passing through a NAT breaks the application’s intended communication model and requires compensating mechanisms.
We are asking the wrong question if we say, “should I deploy IPv6 now”. Someone once told me that IPv6 was here to stay. To my way of thinking, it has not arrived after 25 years.
Let’s look at the situation where we want to merge two large company networks together that both make extensive use of NAT. This becomes more complex than if the two networks were originally using a valid replacement for IPv4 and sadly, that protocol doesn’t exist. While I agree with the notion that the Internet can’t be completely stateless, this doesn’t justify using NAT as middleware. Justifying NAT for the sake of IPv4 life-support is nonsensical.
We should appreciate NAT for its role as a tactical compensating mechanism for IPv4 address depletion, not a a strategic future-proofing scalability mechanism for IPv4. Really what many are saying about NAT is just putting lipstick on the IPv4 pig. Unfortunately, in IT there is nothing more permanent than a temporary solution. Let us not fall victim to this easy psychological trap only because we seem to have collectively painted ourselves into a corner of sorts.