What is IPv6?

• This article explains the history of IPv6 and the format of the protocol header.
• It covers the similarities and differences between IPv6 and IPv4 and how they can coexist and function on dual-protocol hosts and networks.
• It includes current statistics of IPv6 usage and discussed the inevitability and the benefits for all organizations to deploy IPv6.
• This article’s contents are also covered in this video.

 

Internet History

The history of IPv6 goes back to the creation of the Internet.  In the early 1990s, the Internet was privatized and moved out of the exclusive realm of governments and universities into public use.  Based on early indications, it was expected that Internet Protocol version 4 wasn't going to be able to scale in size as the popularity of the Internet grew.  The Internet architects at the Internet Engineering Task Force (IETF) made two key adjustments to IPv4.

First, Classless Inter-Domain Routing (CIDR) allows for IPv4 subnetting, eliminating the classful Class A/B/C address allocation method.  CIDR led to increased fragmentation of the routing tables.  This graph, by Geoff Huston, Chief Scientist for APNIC, shows the growth in the number of entries in the IPv4 default-free zone.

The second adjustment was the introduction of Network Address Translation.  NAT allows organizations to use private RFC 1918 addresses internally and convert those source addresses in packets into public addresses to facilitate Internet reachability.  The use of NAT leads to private address overlaps between organizations.  Now, most IPv4 end-to-end communications are interrupted by several IPv4 NATs along the client-to-server traffic path.

 

IPv4 Address Exhaustion

Regardless of the IETF's efforts, IPv4 address exhaustion occurred.  Nearly a decade ago, all five of the Regional Internet Registries around the world exhausted their supply of public IPv4 addresses.  If an organization needs additional public IPv4 addresses, they can't easily obtain them and must rely on purchasing them on the public address transfer market.  One should ask themselves; do we have enough IPv4 addresses to sustain our IT operations indefinitely into the distant future?

 

Internet Protocol version 4, 5, and 6

The IETF also realized that a new Internet Protocol version would be needed to solve these problems over the long-term.  They set about evaluating several proposed competing options.  The Internet Assigned Number Authority (IANA) maintains the list of Internet Protocol version numbers.  There were IP versions 1, 2, and 3 before version 4.  Version 5 was assigned next to Internet Streaming Protocol (ST2), which was not a direct replacement for IPv4.  Therefore, IP version number 6 was the next version number that was assigned to this new protocol chosen by the IETF to replace IPv4.

 

IPv6 Packet Header

In 1995 the IETF published RFC 1833 - the original IPv6 protocol specification.  In 1998 they created an updated IPv6 specification: RFC 2460.  The final full Internet standard for IPv6 is now RFC 8200.

IPv6 has a completely different protocol header format than IPv4.  IPv6 has a simpler header structure than IPv4, with fewer fields.

In IPv6 packets, the version number is “6”.  The IPv6 traffic class has identical use as IPv4's TOS byte.  IPv6 adds a 20-bit flow-label field for flow-based forwarding.  The payload length indicates the size of the full packet in bytes.  The next-header field indicates the protocol that follows the IPv6 header.  This is the method for how extension headers are used to extend the functionality of the protocol and add optional capabilities such as fragmentation, IPsec VPNs, and segment routing.  The hop-limit performs a similar function to IPv4's Time-To-Live (TTL) field.  IPv6 header includes massive 128-bit source and destination addresses.  Those large addresses provide enough globally-unique non-overlapping addresses for any imaginable network infrastructure

 

IPv6 and IPv4: Similarities and Differences

Both IP versions have a 4-layer protocol layer model.  They are both connectionless datagram delivery protocols.  They are both end-to-end routed protocols using dynamic routing protocols and algorithms like RIP, EIGRP, OSPF, IS-IS, and BGP.  They both have support for one-to-one unicast communications, one-to-many multicast communications, one-to-nearest (or one-to-one-of-many) anycast communications.  They both can be carried over many Layer-2 protocols like Ethernet and Wi-Fi.  They both support TCP, UDP, and QUIC transport protocols and facilitate applications like DNS, HTTP(S), SMTP, etc.  ICMPv6 is like ICMP and both protocols have Quality-of-Service (QoS) markings.

IPv6 is different than IPv4 in many ways.  The most obvious is how the IPv6 header has 128-bit source and destination addresses.  These are far larger than IPv4’s 32-bit addresses.  The IPv6 header has a 20-bit flow label field.  IPv6 does not use broadcast and instead uses multicast for all one-to-many communications.  IPv6 hosts have multiple IPv6 addresses including link-local addresses on IPv6 interfaces and dynamic node interface identifier addressing.  IPv6 uses extension/options headers (e.g. IPsec, fragmentation, SRv6).  IPv6 routers do not fragment packets which require the connection source to do Path MTU Discovery (PMTUD).

 

Dual-Protocol, Dual-Stack

Even though IPv4 and IPv6 are completely different protocols, they can coexist at the same time within a host operating system.  This is referred to as dual-stack, the computer has both an IPv4 and IPv6 protocol stacks.  The physical wired or wireless network infrastructure can transport IPv4 and IPv6 packets within 802.3 Ethernet or 802.11 Wi-Fi frames respectively.  Computers using both IPv4 and IPv6 have the power of being multi-lingual and being able to communicate with anything else on the Internet.

Happy Eyeballs (RFC 8305) is a method whereby         dual-protocol hosts can attempt to connect using both protocols simultaneously, using whichever protocol connects faster.  IPv6 is often faster because those packets are routed via the mode direct path and not backhauled through a translator.

 

IPv6 Usage Statistics

IPv6 has now surpassed IPv4 to become the most popular Internet Protocol.  IPv6 usage has passed 50%, as measured by many different sources.  Therefore, in many places around the world, there are more IPv6 packets/connections/users than IPv4.

Google's IPv6 Statistics have charted the growth in their observed IPv6 connected users over the years as a result of them being an early adopter of IPv6 on their Internet content.

APNIC IPv6 statistics show that more than 30 countries have more IPv6 usage than IPv4.

IPv6 Matrix also keeps track of these IPv6 usage statistics and they are tracking similar global percentages.

 

Summary

• IPv6 is an eventuality and is an inevitability.
• The IPv6 Internet already exists, and it has about two billion IPv6 Internet devices/users.
• Much of your organization's IT infrastructure is ready for IPv6, it is just something you need to intentionally enable.
• IPv6 performs better (on average) than IPv4 on the Internet due to its lack of reliance on Network Address Translation and routing IPv6 packets direct to the destination.
• Companies are adopting IPv6 to modernize their IT infrastructure and avoid technological obsolescence.
• Organizations are migrating to IPv6 to facilitate communication with the broadest range of Internet users, customers, partners, suppliers, vendors, and everyone.

 

Scott Hogg has over 30 years of network and security experience and is president of Hogg Networking (HoggNet.com). Scott Hogg specializes in teaching Internet Protocol version 6 (IPv6) and providing implementation guidance to large organizations. Scott is CCIE #5133 (Emeritus) and CISSP #4610.  Scott is Chair Emeritus of the Rocky Mountain IPv6 Task Force (RMv6TF) and co-author of the Cisco Press book on IPv6 Security.