As we come close to ending this rather long running series on how the Internet really works (because I’m certain you’re about bored of this series, and ready for me to talk about something else!), I’d like to discuss three more topics I think are really important to the Internet’s operation on a day to day basis. The first, herewith, is that “big database in the sky” — you know the one, where all the IP addresses are stored up for a rainy day.
Where do IP addresses, protocol numbers, and things like that come from? All the numbers that “make the Internet go” are managed by the Internet Assigned Numbers Authority (IANA).
In each RFC published by the IETF, you’ll find an “IANA considerations” section. This section outlines any new numbering registries the IETF wants the IANA to create, and how to manage the new numbering space. For instance, IP protocols numbers, well known TCP port numbers, and well known BGP communities are all IANA managed number spaces. This might seem like the “small end of the stick,” though, compared to what IANA is most known for — IANA manages the IP address space.
This means IANA controls which IP addresses are private, reserved for documentation, reserved, etc. — although these reservations are normally made at the behest of the IETF (see RFC 5737 as an example). Once these are taken out of the pool, however, IANA controls where and how the remaining publicly routable IP address space is managed.
IANA was originally funded under a contract through the United States government as a nonprofit organization. There is a small staff that basically manages the systems involved in tracking and assigning these various number spaces, the top level domains, and a few other things — and not much else. IANA doesn’t sell IP address to large network operators directly; the original idea was to prevent even the appearance of profit or “direct lines of control” between the US Government and the numbering resources of the Internet.
When you hear people complain about how the United States controls the Internet, you should keep in the back of your mind that the model was explicitly set up in the first place to avoid that outcome. IANA is being transitioned to a multinational sponsorship model, which may — or may not — be a good thing in the long run. But all this outside the scope of this post’s topic.
IANA is set up, however, to avoid favoritism among the users of the Internet as much as possible, primarily by working through regional registries. Each one represents a block of nations, as shown on this map taken directly from the IANA home page:
Each regional registry is set up under it’s own rules, based on the culture and market which it is serving. As a regional registry needs address space, they ask for blocks from IANA. If the request is properly justified, IANA hands the regional registry a block, out of which the regional registry can “sell” address space to large network operators within their region.
These regional registries have a staff of their own — they are supported by the membership fees of the companies who use address space from that registry. Each registry also offers a unique blend of services, including tools, and information. The regional registries have a conference once a year — normally collocated and run in conjunction with a Network Operator’s Group (NOG — covered in a future column) at which members can present and learn about the direction of the Internet in their region, best practices, etc. Each regional registry web site is a treasure trove of tools and information — for instance, APNIC keeps track of all sorts of statistics in the global routing table. The links to these sites are here for easy reference and exploration —
’til next time — let’s try to keep the jitter down, shall we?