Paper review: IP Next Generation Overview

Reviewer: Michael S. Liu

The arguments for the IPng in this paper are numerous. It does a great job of explaining those benefits in a clear fashion and help to make the success of this new protocol more likely. Hinden outlines the history, specifics, and benefits of the protocol and this is a very effective method for aiding its rapid adoption. It looks like this protocol should be rapidly adopted and implemented though it would be interesting to see a paper outlining criticisms of the protocol. The main reasons that the protocol will succeed are outlined clearly in the paper as:

Though these reasons are all important, I feel the most important for its adoption will low startup costs, minimal upgrade dependencies, and incremental upgrade and deployment. This type of adoption strategy makes the most radical infrastructure changes easiest to bear and most economically feasible. The adoption of the current internet came about in the same manner and if the upgrade of the current IP structure follows suit, it will have a much stronger chance of success. Once the new standards are adopted, in place, and actively utilized, the most drastic but beneficial differences will already be in place and we will begin to reap the full benefits of such an evolutionary and revolutionary shift.

The main need for IPng was of course to make the address space increase from 32 bits and 128 bits. This need is going to be more apparent as we head into the continuing growth of the Internet and especially when we move to place all household appliances onto the net. It will no longer be enough to assign an IP to each person but a whole range of IPs will be required for each household. 2^128 will be definitely adequate as it is comparable to the number of particles in the universe. Even in the most pessimistic estimates, it is 1,564 addresses/square meter, which is still wholly adequate. It is best that we made enough preparations for this in our first correction so that we wholly eliminate Y2K like scenario from ever happening in the address space area.

Another interesting improvement is the ability to distinguish packets based on the needs of the data stream. This makes adjustments for high-bandwidth applications as well as other specialized needs as we build in more functionality into the net. This is an interesting improvement because it builds in a few of the benefits of stratification into the net though it was meant to be a very simple universal system. We may get the benefits of ATM after all with the compatibility of IP which made the Internet a success in the first place.