Depleting IPv4 addresses: Is it time to start transitioning to IPv6? - part 2
In my previous post , I was discussing some of the probable solutions to tackle the IPv4 address depletion problem. In this blog, I would like to list down some of the most popular alternatives being adopted and try to arrive at a best fit.
a .Carrier Grade NAT (CGN): Traditionally enterprises have used NAT (Network Address Translation) as a mechanism to allow multiple internal "private" machines to share a unique public IP address. This 'blankets' the enterprise network from the internet and provides a layer of security. The same concept is replicated in CGN on a larger scale, where the ISP assigns a single public IPv4 address to multiple clients, and the customers in turn share this address between the systems in their local network. Even though this might provide a temporary stop-gap solution, in the long term might not be scalable and result in increased levels of complexity and overhead in managing the networks.
b. Purchase additional IPv4 addresses: By the looks of it, this doesn't seem to be a very encouraging alternative. The lack of IPv4 addresses has given rise to a vibrant market for trading addresses and efforts are on to put in place policies for legitimate address trading. The fallback is that, there is a possibility that organizations having more IP addresses than they need, can hoard the addresses. Again, this would only offset the crisis till a more viable long term solution is available.
c. Migrate to IPv6: During the early 1990's when it was realized that the IPv4 would eventually run out, work was started to develop a new version of the IP protocol and in 1998 IETF(Internet Engineering Task Force) came out with the first version of the new IPv6 protocol. An IPv6 address is 128-bit long and approximately 3.4 x 1038 addresses. To simplify understanding, we can do with an analogy. If we assume that total IPv6 address space is the size of the earth, two IPv4 addresses spaces would fit inside a single tennis ball! So this is literally a limitless supply of IP addresses.
Running through some of the parameters like capex, opex, scalability, flexibility, long term growth and extensibility, NAT (Network Address Translation) appears to be the least appealing, probably only scoring as far as capex is concerned. Evaluation of the second option - 'Purchase of additional IPv4 addresses' doesn't tip the scales on either side after taking the parameters into consideration. The only option which fits the bill now is the IPv6 Migration.
Since early IPv6 days, network equipment vendors had started work on incorporating support for IPv6 in their product suite. The major players continually released newer versions of their products with built-in IPv6 support. But the end users/enterprises were mostly unaware and complacent. Unless there is a compelling reason to shift, most of them would rather stick to easier alternatives. This is what has happened, and there is no real demand for IPv6. If the Equipment vendors are IPv6 compliant but the service providers/enterprises that deploy this equipment are not ready to migrate, it does not make sense at all. It would be like fighting for a lost cause unless there are collective efforts from all quarters. So it is pretty much like a vicious circle, with each waiting for the other to make the move, the IPv6 implementation taking a beating in the long run.
In the next blog, I would wish to explore more on the common IPv6 transition techniques and methodologies.