Top Challenges Network Engineers Face in Migrating to Faster Networks
The race to implement Internet Protocol version 6 (IPv6), the addressing scheme that gradually is replacing the almost exhausted IPv4, has been a marathon, not a sprint. Indeed, the effort has been ongoing for about a decade. And while progress has been made, the deployment of IPv6 is by no means near completion.
Comcast always has been in the forefront of these efforts. The cable operator has been an evangelist on the need for IPv6 and the common sense of rolling it out. It’s a hard sell in many quarters, however. IPv6 implementation is like a flu shot: Its benefits are mainly to prevent something bad from happening. It is not a perfect analogy; a lot of new things are possible with IPv6. But the point is that IPv6 only generates revenue indirectly.
Earlier this month, many sites, including NASDAQ, reported that the Internet Society (ISOC) validated Comcast’s claims that it is the biggest user of IPv6 in the world:
ISOC’s latest report stated that Comcast commands a massive 16.4% IPv6 score, moving ahead of its closest telecom rivals AT&T Inc. ( T ) and KDDI Corp. of Japan.
The story says that Comcast provides IPv6 capabilities to 75 percent of its service area and will blanket its entire footprint by early next year. Other cable operators pale in comparison. Time Warner Cable commands only 3.4 percent of the traffic and Liberty Global only 2.2 percent.
The numbers in the ISOC assessment of Comcast were echoed by the operator itself in a blog post written by Comcast Fellow and the MSO’s Chief IPv6 Architect John Brzozowski. Light Reading added that the next step for Brzozowski and Comcast is to actually get subscribers to use services supported by IPv6:
With IPv6 traffic on Comcast’s network still relatively low, he’s looking to boost the traffic numbers significantly next year, in part by working with the consumer electronics industry to build IPv6 support into more connected TVs and other new smart CE devices in the home.
In the blog post, Brzozowski lauded vendor partners ARRIS and Cisco for the gear they created. Enterprise Networking Planet reports that David Ward, Cisco’s CTO and chief architect, said at the vendor’s Tech Radar Event during the first week of December that the Internet is capable of supporting the billions and billions of new addresses IPv6 will make possible and pointed to growth in usage of the addressing scheme in 2014 and beyond.
A trend to watch is the marriage of IPv6 and software-defined networks (SDNs). SDNs are an approach in which the nature of the networks (i.e., how the underlying hardware is provisioned and used) is controlled in a centralized manner. The current approach, in which this intelligence is co-located with the router and switch hardware that sends the packets on their way, is inefficient and cumbersome and, thus, will fade. Patrick Hubbard, the senior technical product marketing manager at SolarWinds, wrote a commentary at SearchSDN on the relationship between SDNs and IPv6.
The implementation of SDNs will be complex. However, once they are in place, day-to-day operations are simplified. The bottom line will be that the ability to assign the correct address in a certain instance via the old or new scheme will be baked into the system:
This kind of automation and simple management will change the way we look at IPv6. With IPv6, we focus too much on the differences with its ancestor, IPv4. But over time, with SDN we’ll come to think of resources by name first, and the mapping to physical addresses will move under the covers of directory services and controller databases.
Networks are by definition complex. Massive changes, from IPv4 to IPv6 and from the current management structures to SDNs, magnify that complexity in the short term. The long-term vision, however, is for more homogenous, responsive and higher-capacity networks. The next couple of years will determine whether that vision can be realized.
