No-IPv4 or IPv4less or Simply IPv6-only?
Nomenclature in computer and networking technologies can often be confusing and sometimes even funny.
We IT people have crazy ways of naming things and we frequently use acronyms and initialisms (or alphabetisms) to describe multi-word concepts. Just in case you forgot the difference between these two terms: An initialism is where we say the first letter of each word individually, such as DNS, DHCP, or IPv6. An acronym is where you pronounce the first letter of each word as a completely new word, such as NASA, BIND, NAT or ISATAP. And then there are times when we break these rules and use other letters to form the initialism or acronym, such as NNX or IXP (where the X represents the word exchange). Then there are times where we use a half-initialism and half-acronym, such as DDoS (e.g., pronounced “dee-doss”. Furthermore, FAQ, can be said as either an initialism or an acronym and everyone knows what you mean.
Sometimes acronyms are derived from computer functions, such as ROT13. ROT13 is pronounced “wrote thirteen” and is a method of rotating all the ASCII text characters by 13 (out of 26) characters to create a simple cypher to indicate spoiler alerts on Usenet news feeds. Try your favorite ROT13 converter to decipher this “secret” message (VCi6 vf Orfg).
The Urban Dictionary is filled with funny terms derived from our use of technology like ROTFLMAO, BTW, BRB, IMHO, AFAIK, NSFW, YMMV and meme. Speaking of which (IDNTIMWYTIM):
IT professionals can also be very particular about how these terms are pronounced. For example, the venerable text editor vi is pronounced as an initialism, but its improved clone Vim (originally released for the Commodore Amiga) is pronounced like an acronym. There has also been much confusion about how to pronounce the IPv6 transition tunneling mechanism Teredo, but it is supposedly named after the shipworm, whose taxonomic genus is Teredo.
Sometimes IT people take these acronyms and initialisms and turn them into verbs. You might hear someone say “I vi’ed the file”, “the server DHCP’ed for its address”, “the firewall NAT’ed the connections”, and even “we VLAN’ed the network switch”. Or “DM me when you get a moment” referring to the Direct Message feature in Twitter.
Sometimes IT folks take a common word and use it in a different context. An example of this would be the term cookies. And then again we completely invent our own words like BOGON, “133t”, and “n00b”.
In recent years, there have been some technologies that have been described by what they are not rather than what they are. It is not that these people are preaching neo-luddism and are totally against technology innovations, but rather they have a strong opinion about a particular technology.
One example of this is the term “NoSQL”. NoSQL refers to a database that does not use the Structured Query Language (SQL) that has been a 30-year old method of interacting with a relational database. The NoSQL proponents are proud of their use of unstructured databases that perform auto-sharding, which breaks from the tradition of using SQL. To highlight this distinction, the term for this whole set of new databases is the antithesis of the previous-generation technology.
Another example of such an anti-term is “Serverless” computing. Serverless architectures are a model of cloud computing where a cloud consumer simply deploys their code into the cloud service provider’s system and runs it on demand. The customer doesn’t need to concern themselves with virtualized servers, or even software containers for that matter, their code just runs. The cloud consumer only pays for the amount of memory their code used and the time it took their code to run. The serverless proponents are proud of the fact that it is just their code running and they don’t have to concern themselves with the underlying server infrastructure. Some people don’t like how the term “serverless” is an anti-term and uses a negative to describe something positive. Some are lobbying to have the term changed to Function as a Service (FaaS), but the serverless term has caught hold and it might be too popular to change. Although, I must admit, “V4Less” has a concise ring to it.
Should we use No-IPv4 or IPv4less?
If we were to apply this anti-terminology to IPv6, then would these IPv6-only networks be called No-IPv4 networks or IPv4less?
Many in the IPv6 community are enthusiastic about the benefits of “the new hotness” that is IPv6 and consider IPv4 the “old and busted” Internet protocol. IPv6 proponents often preach the benefits of returning to the purity of the original end-to-end principle of computer networks, and have a natural revulsion for the use of NAT. IPv6 proponents often are so polarized against IPv4 that they might even prefer to think of IPv6 as the “anti-IPv4” protocol.
There are many in the IPv6 community who would prefer to run an IPv6-only network. We in the Infoblox IPv6 Center of Excellence (COE) have written frequently on the trend towards IPv6-only network environments. We have written about how IPv6 can actually be faster than IPv4 and why organizations might consider an IPv6-only deployment. We have written a 3-part blog series that covers how an ISP can use an IPv6-only core to connect subscriber legacy IPv4 devices (Part 1, Part 2, Part 3). We have written about how enterprises may move from IPv4-only to dual-stack to IPv6-only. We have written about the how to build the enterprise business case for IPv6-only and how IPv6-only networks are more realistic that you might think. We have also emphasized how organizations may move from dual-stack to IPv6-only.
Cisco has recently announced how they have dedicated an entire building in their San Jose California campus to be an IPv6-only environment. Cisco’s Building 23 will be a testing ground for using a single network protocol. Microsoft has also been working on IPv6-only environments in their infrastructure to test applications and single-protocol use-cases. You are also likely to hear this IPv6-only message repeated at the upcoming 2017 North American IPv6 Summit conference at LinkedIn on April 25-26, 2017.
A Word on IPv6 Terminology
When it comes to IPv6, there is a whole new set of terminology. As defined in Section 2 of IETF RFC 2460, an IPv6 “node” is any device that implements IPv6 and a router is any “node” that forwards IPv6 packets. An IPv6 “host” is an IPv6 node that is not a router, and is an end-system on an IPv6 network.
With IPv4, the Address Resolution Protocol (ARP) is used to create a mapping of IPv4 address to layer-2 MAC addresses (referred to as an ARP cache). With IPv6, the Neighbor Discovery Protocol (NDP) is used to create the binding of IPv6 addresses to layer-2 MAC addresses (referred to as a neighbor cache).
IPv4 packets have a Time-To-Live (TTL) field in the header to help prevent packets from endlessly looping through the network. IPv6 packets have this same function, but the header field is referred to as a “Hop-Limit”. There are many examples of how IPv6 is functionally equivalent to IPv4. However, we should remember that these two protocols have very different address formats and header formats.
Because Classless InterDomain Routing (CIDR) and the process of performing address subnetting are IPv4 terms, they are not applicable to IPv6. All IPv6 addresses use “slash” notation (not to be confused with the popular rock guitarist). IPv6 network addresses are referred to as prefixes and have a prefix length.
Picture Source: David Neal @reverentgeek https://twitter.com/reverentgeek/status/789135336437800960
IPv6, as with most IT concepts, comes with its own set of terminology and nomenclature. We can relate the IPv6 terms to the IPv4 terms we have been familiar with for decades.
More organizations will realize that dual-protocol is a transition strategy that eventually leads to running a network solely with IPv6. Even though we can refer to running a single-protocol IPv6 environment, we should refrain from using the anti-technology terms of No-IPv4 or IPv4less.
Most of the time “you are what you is” as Frank Zappa so eloquently put it. But other times you could be the opposite. When it comes to computer networks, and more specifically IPv6, we should stick with the former and simply refer to the concept as IPv6-only.