IPv6 Is Tactical, NDN Is Strategic

My last assignment with the National Science Foundation (NSF), in 1995, was IPv6. Even 20-plus years ago, we understood that IPv6 was a tactical protocol designed to correct IPv4 limitations such as the fixed number of IP addresses. As enterprises exhaust their IPv4 networks, they should look past IPv6 as the answer to their network addressing challenges. IPv6 has the following limitations:

  1. Additional Overhead -- For voice traffic, Internet of Things (IoT), and other network traffic that has small packets, the additional overhead associated with IPv6 increases bandwidth requirements. IPSec/v6 tunnels further exacerbate this problem, with a G.711 call having a 200% network overhead.
  2. Minimal Internetworking QoS -- Going from IPv4 to IPv6 and between NAT boundaries causes the loss of QoS and routing policy information
  3. Zero Application Awareness -- It's still a packet-by-packet routing technology and is unaware of network session state
  4. Lack of Security -- The Internet Architecture Board recommends that all future protocols support end-to-end encryption

Named Data Networking (NDN) started in 2010 as an NSF research project to create the architecture for the future Internet. NDN changes the paradigm used by traditional networks, moving away from IP and using numbers for address assignment and routing. NDN defines a network to transport data containers between two endpoints, or nodes, with unique names (similar to URLs). The NDN scheme allows for blocks of data to be stored, digitally signed, and transmitted across nodes, the names of which higher-level applications can understand.

IP vs. NDN protocol stack

The killer application for IPv6 is larger address space, but at a price of additional overhead. Yes, IPv6 has some other advantages, including multicasting, but none that are driving enterprises. With only about 12% of IPv4 addresses in use, the IPv4 address problem is more of an allocation challenge than lack of inventory. We can limp along in the IPv4 world for years, as we have done for the past 20 years, until something truly better comes along. NDN has the potential to be that truly better approach as the type of traffic on the Internet changes (lots of video and content), and the needs of the Internet change (more security).

Whether NDN or a variation of it replaces IPv6, what's becoming clear in software-defined networking (SDN) is the concept of IP address abstraction. So, while an endpoint still has an IP address for now, this address can be associated with a name that's readable by humans and applications. With IP address abstraction, it's possible to support duplicate IPv4 addresses and networks, allowing an enterprise to use RFC 1918 private addressing such as 10.x.x.x multiple times. It could do this to support new IoT devices, for instance.

The drivers to replace IPv6 are:

  1. Zero Trust Networking -- Not allowing a network session to be established prior to authorization and permission
  2. Growth of Video -- As video grows to 95% of the consumed bandwidth on networks, it leads to the need for a protocol that's friendlier for storing, managing, and transmitting video
  3. Edge Computing -- The growth of augmented/virtual reality and content at the edge of networks requires moving routing from the core of the network out to the edge, where every millisecond of delay counts
  4. 5G & IoT -- 25 billion devices talking to four billion users and two billion applications requires internetworking and security at scales for which IPv4/6 weren't designed. Location independence is an important part of this as users roam between networks.

Understanding the need for NDN, Cisco acquired the PARC NDN platform, called Content Centric Networking (CCN), in February 2017. Cisco is branding its version information-centric networking (ICN), which it plans to open source within the Linux Foundation's FD.io (fast data - input/output) community. Quoting from Cisco's blog announcement on its plans:

Cisco's acquisition of PARC CCN has the objective to foster convergence of various dialects of ICN (CCN and NDN) into a single harmonized version of ICN, promoting wider and faster adoption of ICN-based solutions required to solve future networking needs.

Gartner places NDN at the very beginning of its hype cycle, and predicts that it's 10 years away from mainstream adoption. While I agree that NDN is just emerging, I believe the time to mainstream adoption will be quicker due to the demands in the market and the move to SDN.

Related content: