Network Documentation

[#abrnetxt]: Abbreviating network documentation
Stop using IPv4 subnet masks

Prefix lengths (using “CIDR notation”) are faster to type and read, and are a better habit to get into (realizing that prefix lengths are what get used with IPv6). There may be some cases where subnet masks are more appropriate, such as if dealing with software that requires a subnet mask as input, or perhaps when also dealing with a “wildcard mask” when using Cisco's IOS. However, the majority of the time it is going to be faster to use /24 instead of the subnet mask 255.255.255.0.

This practice may take some time to transition to for those who are used to writing out subnet masks. Such a transition may be of very limited use to somebody who is just documenting things for a one-time project, and who has an understanding of subnet sizes based on memorized subnet numbers. However, for anybody starting out in the IT field and expecting to be heavily involved for years and even decades to come, this transition is likely a good idea since prefix lenghts are how these subnet sizes are written with standard notation when dealing with IPv6.

The only time that subnet masks would conceivably be more useful would be in the rare case when the subnet mask being used is not one of the subnet masks that directly corresponds to a prefix length. However, this would certainly be an exception to standard practice, because such an approach is uncommon. Not only would such an addressing method provide little to no benefits in most cases, but such an approach (especially if implemented improperly) might not be fully supported by some equipment, including computers running software that may not fully support such a non-standard approach. So, this exception to the rule (of using prefix lengths instead of subnet masks) would not be a commonly recommended practice.

[#adypartl]: Using partial addresses

When filling out internal documentation, consider using partial addresses, particularly during the early stages of the documentation process. This may also be useful if documenting something which it is likely that nobody will ever read (but which should be documented just in case someone does read it), such as a description of some work that was done. (If the work was successful, chances are the documentation won't be needed. If the machine does need to be worked on again soon, possibly even for another reason, it may be helpful to have a record of some details on what exactly was done.)

(e.g. if the IPv4 range is 192.168.245.0/27, a partial address here of .3 refers to an IPv4 address of 192.168.245.3). Partial addresses are faster to document, which makes the documentation process go more quickly. Also, if the first part of a subnet's address needs to change during the early planning, then documentation that is using partial addresses may not have any text that needs to be updated to reflect the change.

For actually documenting these addresses, be absolutely sure that the subnet is documented in a prominent locatation, easy to see when working with partial addresses. It may even be worthwhile to just convert all of the addresses to full addresses when recording in some long term documentation that is expected to be used down the road. If the actual long term addresses (or a piece of those addresses, like the subnet being used) need to be looked up later, that requirement will cost time later. Worse yet, if the partial address is expanded differently at different times (perhaps because somebody working on the network later is thinking of a different subnet), then some machines may accidentally be configured to use different subnets. This may result in communication not functioning as intended.

As an example, if documenting the critical details of the IPv4 configuration of a machine, the note may be abbreviated to something like “192.0.2.51/24 DG=.1” The term “DG” is meant to refer to the “Default Gateway”. Those who know that abbreviation should be able to figure out that the default gateway's full address is presumably 192.0.2.1. (Keep in mind that the default gateway would need to be on the same subnet. If there is a reference to a partial address for something like a DNS server, chances are that it will be on the same subnet. However, it wouldn't strictly need to be. A DHCP server usually is on the same subnet, since DHCP Relaying to a different subnet isn't super common.)

[#skelsys]: Skeleton document for system documentation

The skeleton document for system documentation may provide information that is useful to document about an individual computer. There is also a skeleton document for system documentation (text file with blanks) which is available.

System Address Usage

See: System Address Usage which contains some samples of how to number various systems on a network.

[#lanipskl]: Making documentation of a LAN
See: Making documentation of a LAN

Additional resources: network address planning