Why isn’t IPv6 taking off?
Thu 8 Jan 2015
Mike Palladino, director of IP infrastructure and operations at Internap, asks why the adoption of IPv6 is still in its infancy.
Routing IPv6 packets is well-supported within most modern hardware, but many organisations are struggling with all of the ancillary things that have to happen to facilitate IPv6 — retraining their NOC, rebuilding their management, monitoring, and troubleshooting tools to speak both IPv4 and IPv6, developing IPv6 operational experience and so forth. According to Google, on December 9, 2014, 4.29% of all of its traffic was IPv6. With IPv4 addresses near depletion, why is the adoption of IPv6 still in its infancy? There are three key reasons:
- Squeezing blood from the IPv4 stone is easier than making large design changes. The initial implementation of IPv6, or converting to dual-stack IPv4/IPv6 will always take a backseat to fighting today’s fires.
- A burgeoning IPv4 broker market has sprung up, allowing private resellers to satisfy IPv4 needs for IT organisations in the likely case that the Regional Internet Registries (such as ARIN, RIPE, APNIC, and others) cannot assist.
- Hardware limitations that potentially discourage the use of IPv6, as well as expensive and time-consuming hardware upgrades that are necessary to support IPv6.
As mentioned above, migrating to IPv6 requires substantial network and hardware design changes, even though most routers are able to route IPv6 packets. In addition, the operational tasks that come along with supporting an IPv6 network are also a challenge. There are a few critical issues in this area that need to be considered:
- Do all of your IT organisation’s applications support IPv6? If not, your monitoring system, troubleshooting tools, IP management software, application suite or other software packages might require lengthy deployments, replacements or development.
- Is your organisation’s NOC adept at troubleshooting IPv6 issues? If not, that means training across a wide range of staff and skillsets to learn new commands, use new tools and troubleshoot things a bit differently.
- Do your documentation, process manuals and event maintenance scripts take IPv6 into account? If not, you’ll likely need to refresh it, which takes time and effort.
- Does all of your hardware natively support IPv6 end-to-end, including your routers, firewalls, load balances, VPNs and servers? If not, you might be looking at costly, time-consuming, and customer-impacting replacements of legacy equipment.
As a result of the above, organisations are avoiding these migration headaches, preferring to squeeze more mileage out of the remaining IPv4 address pool via the broker market. As IPv4 space gets scarcer, the “dollar per IP” cost to get additional IPv4 space from a broker will continue to increase. Eventually, there will be an inflection point where the cost of brokered IPv4 space will be painful enough to a business that it will outweigh the operational effort to migrate to IPv6; however, we clearly are not there yet based on the amount of business that IP space brokers appear to be doing.
Finally, most IT organisations recognise that, since IPv6 routes take up twice as much space in a router’s memory as IPv4 routes, a migration to IPv6 can potentially accelerate the need for costly hardware upgrades. However, as more organisations de-aggregate the IPv4 routing table by avoiding a migration to IPv6, the same limitations still exist – it’s very much a case of, “pick your poison.” Either your hardware limitations are reached by further IPv4 de-aggregation or a migration to IPv6, but regardless, hardware upgrades are looming for companies that have not had to refresh their hardware in some time.
As these hardware limitations are reached, businesses are going to need to do wholesale hardware upgrades, or they will have to take less than a full BGP routing table from their provider. Taking less than a full table from the upstream provider will impact the granularity of routing control and the level of insight into what’s going on with the full Internet table, which is definitely a step backwards. One thing is clear – over the next few years, millions of routers will hit their physical limits due to IPv4 de-aggregation and IPv6 adoption, and IT organisations will be faced with spending enormous amounts of money to move to next-generation platforms.