As I write this I’m returning from this year’s FutureNet conference in Boston where much of the discussion centered around the twin challenges of dealing with depletion of the IPv4 address space and the growing concern that IPv6 adoption will exacerbate scalability issues within Internet routers. While there was much disagreement about the best way to address these issues, the majority of speakers agreed that despite longer-term route scalability concerns, a migration to IPv6, whether we like or not, is in our immediate future.
Why? Quite simply because the immediate concern among Internet architects is that we’re running out of addresses. Depending on whose estimate you believe, we’ve got another one to two years before Internet service providers have no more addresses to assign, meaning that if you want a new block to support a new facility, or your provider wants to expand its mobile offerings, you, and they, will find that they lack the IPv4 addresses to meet demand.
There are a few options for short-term fixes. One is to create a market for excess IPv4 address space to enable those who have extra space and who today have no incentive to give to sell it to whoever needs it. Another is to increase the use of network address translation to hide private networks behind small numbers of shared public addresses. Both of these create additional concerns: who would manage the market to ensure that large entities don’t gobble up all available address space? And, there is the potential for carrier-grade NAT to disrupt the Internet architecture that’s based on smart end-points and a dumb core.
Meanwhile, most major Internet service providers have begun to implement IPv6 within their networks to solve the address shortage problem. But IPv6 migration creates new and possibly more severe problems as Internet routers already taxed with maintaining rapidly growing route tables now must cope with the addition of IPv6 routes (most providers are using a “dual-stack” approach of running both IPv6 and IPv4 at the same time), and inevitable continued fractioning of the IPv4 address space. Both of these issues point to a future where Internet reachability issues will increase, regardless of available bandwidth. Unfortunately, the topic of Internet scalability is receiving scant attention in the media. Enterprise IT architects still don’t see any business case compelling them to adopt IPv6 internally, or lean on their providers to address route table scalability. For most the Internet continues to operate as a black box that has always worked, always scaled, and always grown to meet emerging application demands.
But that perceived reality is poised to change. Enterprises will run into address shortages as their providers run out of IPv4 space, then they’ll deal with reachability and performance issues as route scalability problems become pervasive. As UC increasingly crosses company boundaries, addressing and routing issues could threaten the ability to reliably deliver such services as inter-company video, voice, or presence federation.
In the short term it’s incumbent on any IT manager to get up to speed on Internet scalability issues and prepare for the potential for IPv6 adoption at least at the service provider edge in the next few years. IT leaders should begin to evaluate internal applications to determine options for delivering IPv6 support, and they should work with their service providers to understand their own approaches for meeting addressing and routing challenges going forward.