Written by: Daniel Smith 7/10/2015
If you haven’t heard, ARIN, the governing body for IP address allocation in the US, recently ran out of IPv4 addresses. Some folks are alarmed while others are more pragmatic; after all, we’ve already identified the solution. Though the end of IPv4 will likely take decades and IPv6 has been around for quite a while, IPv6 adoption by ISPs will certainly accelerate as the secondary market for IPv4 addresses can’t keep pace and IPv6 becomes necessary.
A lot of the current buzz around the Internet-of-Things, or IoT, surrounds what I would categorize as thin-sensors. A lot of sensors contain just enough power and electronics to do their sole job and communicate with a parent device. Be it an Apple Watch, FitBit or Metromile Tag, all of these devices rely on a smarter device within relatively short proximity to be the brains of the operation. This is akin to the thin-clients of the 1980s–purpose built and delivered as cheaply as possible with a smarter, more powerful device backing it up. While I would consider these devices a subset of IoT, possibly labeling them consumer M2M, IoT by its own name suggests that each device be internet-connected. The current influx of devices are not directly connected and therefore, do not require a unique, public IP address. This will very likely change.
Gartner suggested that 4.9 billion “things” will be connected in 2015 alone and the IPv4 address space has just under 4.3 billion IP addresses. As the demand for direct-to-Internet connected things grows, so will the demand for IP address space. The same Gartner report also suggests the number of connected devices will increase exponentially to 25 billion by 2020. IPv6 needs to be a key focus for the entire IoT industry as its viability depends on reliable, direct connectivity for each device. IPv6 is the only long-term IoT addressing solution that removes smarter, close proximity devices and gives unilateral connectivity to the device itself. Validating IPv6’ long-term efficacy, I once heard someone describe IPv6 as being able to assign an IP address to every atom on Earth. While this isn’t quite true, it gives a vivid image of the magnitude of the IPv6 address space.
Of possibly greater importance than the address space itself are the networks carrying the IPv6 traffic. While networks historically carried and were measured by IPv4 performance, IPv6 is largely untested except by the few who seem to care. The real consumer of IP address space, Internet Service Providers, will likely have to start focusing more of their efforts on sustaining growth of IPv6 traffic and management of their networks while the transition is ongoing. I’m hopeful their progress will be evident in the speed and reliability of their networks.
Though maybe not as critically, there are two industries that will be curiously affected by widespread adoption of IPv6: email senders and DNS. Email senders have an interesting attachment to IPv4. IP address reputation plays a very significant role in suppressing unwanted email and IPv6 makes managing reputation by address an extremely difficult proposition. I’m curious to see how ISPs and their vendors respond. DNS will likely increase in complexity as it tries to make sense of all the IPv6 address space, allowing it to be logically addressed and accessed by humans. I suspect the Address Bar will fade into the background of most browsers, hidden in a tertiary menu used by the few nerds who won’t let it go, replaced by something new… proximity-enhanced perhaps?
Every baby step taken toward the complete adoption of IPv6 is one step closer to more connectivity for more devices. We should hail the end of IPv4, be excited to adopt IPv6, and be thankful that the exhaustion of IPv4 addresses is playing a crucial role in progressing IPv6 as the final solution.