Can the internet really handle mobility?

The base architecture of the internet contains two fundamental and significant flaws. One has been addressed, but one has not.

The internet relies on the basic assumption that every device must have a unique address. That includes every PC, every PDA, every GPS unit... everything. The current protocol for assigning addresses is called IPv4, and it allows for a finite number of addresses. Look at the accompanying graph and you see that we are rapidly approaching the upper limit of available addresses.

A new protocol, called IPv6, is meant to resolve this problem. However, a recent study indicates that less than 1% of IT executives say they are deploying IPv6. Furthermore, those respondents cite government mandates as the only reason they have gone through with the deployment. It is unlikely that IPv6 will be widely deployed until government or market conditions compel organizations to do so, making it very probable that this particular IPv4 limitation will have a significant impact on all of us.

However, even if IPv6 is deployed flawlessly across the world tomorrow, we still have a significant problem, especially as it relates to mobility. Once a device is assigned an address, the internet establishes a path to that device. And the details of that path have a significant reliance upon geography. As a device moves from one region to another -- in an airplane or a ship -- that path will need to be updated. The computing power required to keep track of all of those individual paths, and their changes is virtually incomprehensible.

Unfortunately for all of us, the fundamental design and development of the base internet was "good enough" for the purposes of the last 15 years. However, to paraphrase Jim Collins, we've allowed "good enough" to become the enemy of progress. I'm sure there is no simple solution. But ignoring this problem will invite significant disruption in our global connectivity.

Read the findings from Nemertes Research: Internet Interrupted