I've spent the last several days at the FutureNet conference in Boston, an annual event focused on WAN services such as MPLS, as well as the broader evolution of the Internet.  This year's event featured a round-table discussion by numerous Internet luminaries who all seemed to reach the same conclusion: the Internet architecture is breaking, and we're a short time away from major problems.

Without spending too much time diving into the specifics of BGP, IPv6, and other Internet architecture topics, the basic argument is that we're nearing a two-fold "perfect storm" of Internet challenges.  The first is exhaustion of IPv4 address space, with estimates ranging from 3-10 years or so before we run out of assignable address space.  The second is the number of addresses being routed, with route tables growing faster than processing power.

Nobody had a single "magic bullet" to fix these problems, and approaches such as IPv6 or new routing protocols could take years to implement, and may cause more problems during the transition period.  While prognostications of Internet calamities have been around as long as the Internet, the fact that so many individuals with deep knowledge of the Internet architecture are reaching agreement of the problem (though not the solution) is significant, and is worth noting by those relying on the Internet for applications, commerce, or other related activities (basically - all of us).  It may not be "sexy" to focus on the underpinnings of the Internet, but increasingly we can't ignore it.