This document describes the NETS standard for naming network devices, cables and locations at UCAR. This standard was used for many years before the UCAR Building Numbering and Interior Signage Standard was written. As much as possible, NETS has integrated the UCAR standard into this standard.
In general, there are two ways to refer to a piece of network equipment: by a name given to the hardware, like router4 or PetesPC, or by the location of the hardware, like building 3, room 4, rack 5. Both kinds of information are useful when debugging networks. The device name is used when dealing with connecting a network logically - with software. The location name is used when connecting the network physically.
To reflect the two kinds of information, network devices and cables can have two kinds of labels, "name" labels and "location" labels. The syntax of the labels is different to make it easy to tell which kind of label you're reading. To make the difference more clear, name labels are always printed in lowercase, and location labels are always printed in uppercase.
There are two types of names, host names and network device names. Host names, like fileserver or meeker, are chosen to be easy to remember. Network devices like routers and switches have names that contain information about the device, such as location, type, etc. These descriptive names are useful when networks are being debugged. This section describes both kinds of names.
Names of computers are like fileserver or netsdb. Hosts that have multiple interfaces have multiple addresses, and therefore can have multiple names. When assigning names to addresses in the Domain Name System (DNS), a single host's multiple names may look like:
host-all resolves to all REACHABLE addresses. There is one host-nxx for every interface (whether reachable or not), where the "xx" is the network "number", in most cases simply the third byte of the 220.127.116.11 network numbers. If the network portion of the address extends into the fourth byte, the name should take the form host-nxx-yy where "xx" is the third byte of the network address, and "yy" is the fourth byte. For example, serialr is an interface with the address 18.104.22.168. The network address for this interface is 22.214.171.124, and it's name is serialr-n243-12.
Concerning reverse resolution, host and host2 take precedence over the -nxx form if there is a conflict. Note that host2 would be used only if users needed a second name (in case the first name failed.)
NETS has chosen to use all lowercase for names, because most of the systems that connect to UCAR networks use DNS names, and DNS names are traditionally all lowercase. Using all lowercase also helps to distinguish name labels from location labels, which use all uppercase.
Names of non-computer devices, like routers, follow this general syntax:
Site - Room - Manufacturer and unit number - Device Type
For example, ml-16c-c1-gs.
The same UCAR site names are used on all types of labels, except that lowercase is used in name labels and uppercase is used in location labels . The site names are:
The room field contains the standard name for a room. It is usually just a number, but can contain letters. For instance, 31h is the name of a room at the Mesa Lab. For historical reasons, the computer room in the Mesa Lab is named "mr" in name labels, even though the room is technically number 29 or 32. Name labels are all lowercase, including the room name, even though room names elsewhere might be uppercase.
The manufacturer field indicates the name of the manufacturer.
|apc||American Power Conversion|
|asi||Access Specialties Inc.|
The unit number field specifies a number that differentiates the unit from other units. Unit numbers start at 1, but don't "mean" anything; unit 1 isn't "primary", wasn't necessarily installed before unit 2, just because there's a unit 3 doesn't mean there has to be a unit 2, etc.
The device type field is used to define the function of the device, independent of the manufacturer.
|acc||access card controller|
|acr||access card reader|
|ap||wireless access point|
|gs||gigabit Ethernet switch|
|gw||gateway (a.k.a. routers)|
|sar||stand alone repeater|
|scc||security camera controller|
|ups||uninterruptable power supply|
|vg248||Cisco analog phone gateway|
|vpn||virtual private network|
Name labels are the physical labels attached to network devices or cables. On devices, they provide the name of the device. On cables, they provide the names of the devices at the two ends of the cable. On cable labels, the labels indicate the devices at the very ends of the cable connection, even if there are intermediate patch panels. Name labels on equipment should appear on the front and the back of the equipment.
Name labels contain the host or network device name as described in the "Names" section, preceded by a "n:" or "<:" , and possibly followed by an optional port designator in parentheses and/or an optional network name or number in square brackets.
labels can appear on devices, device ports, and cables. When
they appear on a cable, they appear next to a location label
(see "Location labels"). Unlike
location labels on cables, name labels on cables are used only at
the endpoints of each equipment-to-equipment connection.
Name labels on devices and device ports have only one line of information. Name labels on cables have two lines, one specifying the name for the device at the near end of the cable and one specifying the name for the device at the "other" end of the cable.
A name label can contain one or two lines of information. Each line in a name label follows this syntax:
n:site - room - device type - manufacturer and unit number (port) [network]
<:site - room - device type - manufacturer and unit number (port) [network]
This follows the syntax of names described earlier, with n: or <: added to the beginning, and optional port or network fields added to the end. The port and network parts are optional, but recommended. For example:
The last label refers to the le0 port on meeker, with IP address 126.96.36.199.
The port field is the port on the device, surrounded by parentheses. This optional field specifies the name of the port as used by the device's software.
This port name will usually be different than the port name that appears on location labels.
For instance, the name of a port on an SGI machine might be "ae5", and the location label for the same port might indicate the port using "B/0".
The network field is a network designator, surrounded by brackets. This optional field designates a network name or number. It can be a full IP address, a partial IP address, or a VLAN or ELAN name. A partial IP address might be ".32.16", indicating the IP address 188.8.131.52, or ".7.", indicating the 128.117.7 subnet.
This is an example of a Cisco Catalyst 6500 connected to a patch panel, so the name label on the patch panel indicates the 5000, not the patch panel.
There are two types of name labels: device name labels and cable name labels. Device name labels are used on devices and device ports. Cable name labels are used at the endpoints of each equipment-to-equipment connection, and show the equipment nearest the label and what's at the other end (<).
(Link to Cheat Sheet)
This section describes the labeling system used to give unique location information to network devices. Network devices are defined as equipment, cables, and connection devices. Equipment is defined as routers, repeaters, workstations, PCs, etc. Cables is defined as all cables, including twisted pair, fiber, telephone, and serial lines. Connection devices are patch panels and telecommunications outlets (TOs) (wall plates).
Location labels are used to identify many kinds of things, including cables, machines, patch panels, wall plates, racks, single ports on network devices, etc. For some of these items, such as a rack, a simple one-line label is sufficient. For other items, two lines are required. For example, labels attached to cables must identify the location of both ends of the cable, regardless of which end the reader is holding. These differing requirements mean that different types of labels are needed in various places in the network.
A location label is composed of one or two lines. At least one line is found on every label, and describes the location of the local network device. In other words, the local area where a person is observing the label. This line always starts with an "L", for "Local" network device. A second line may exist on the label to provide more information about the labeled entity. It doesn't matter if the second line is above or below the line that starts with "L". On cables, the second line starts with an arrow (facing left or right, it doesn't matter), and identifies the location of the device at the opposite end of the cable. On connection devices, such as patch panels or wall plates, the second line starts with a "B", and identifies the location of the device on the other end of the cable that is on the back of the connection device.
L = The local network device itself (used on all labels)
B = Back, the location of the other end of the back cable (used on wall plates, patch panels)
→ or ← = Arrow, indicating "the far end of this cable", the opposite end location (used on cables)
Each line of location information is composed of four groups of information:
The "-" character is used to separate the groups as shown below.
L:Site - Room and Sub-location - Connection Device - Card and Port
B:Site - Room and Sub-location - Connection Device - Card and Port
<:Site - Room and Sub-location - Connection Device - Card and Port
The same UCAR site names are used on all types of labels, except that uppercase is used in location labels and lowercase is used in name labels . The site names are:
In the following example, the site is the Fleishmann Building. The rest of the label specifies room 5, rack 3, equipment 3, card A, 10base2 media, port 3.
The room and sub-location field specifies a room and a location within the room. If the contains only a few devices, the sub-location field is not needed. The syntax is
The room field contains the standard name for a room. It is usually just a number, but can contain letters. For instance, 31H is the name of a room at the Mesa Lab.
The sub-location field is optional. When a room contains several devices, the field is used to identify a floor tile, rack, or wall-mounted board within the room. The field consists of a letter that specifies the sub-location type, followed by a number or letter/number code for the specific object. The sub-location types are:
Codes for floor tiles are usually defined only for large rooms such as the Mesa Lab computer room. These codes consist of an X/Y pair like D44, where the X coordinate is written with letters and the Y coordinate is written as a number.
In the following example, the Room and Sub-location group specifies room 20, rack 12. The rest of the label specifies the Foothills Lab building 3, patch panel 3, Category 3 media, port 2.
In the following example, the Room and Sub-location group specifies room 29, floor tile D14. The rest of the label specifies the Mesa Lab, equipment 1, single-attach FDDI media, port 8.
The connection device group consists of a letter that specifies the connection device type and a number to identify an individual device. The connection device types are:
Equipment refers to routers, repeaters, hubs, concentrators, and other network boxes.
Note: Modules of a stackable repeater unit are treated as cards making up one piece of equipment, because they have one IP address, and are accessed as one device by network management software.
In the following example, the connection device group specifies equipment 1. The rest of the label specifies rack 1 in room 47 in the Mesa Lab, single-attach FDDI media, and port number 8 on the equipment.
The Card and Port group specifies the card (if any), the media type, and a port. The syntax is
The Card field is optional. It is always a single letter. The letter Y is special. It is used to label unique function ports, such as console, admin, or diagnostic ports.
The Media Type field is always used on lines that start with an "L", and is never used on lines that start with a "B". On cable labels, it is optional on lines that start with an arrow, since these lines always appear with an "L" line. The "L" line shows the media type, so it isn't necessary to repeat it on the arrow line. The media types are:
The B3 media designator is used on the bundled CAT 3 cables that connect patch panels to Ethernet cards on Cisco Catalyst 5000 switches.
The BS media designator is used on the bundled serial cables that connect patch panels to terminal servers.
The DF, MF and SF media designators are used only on the ends of FDDI MIC cables.
The Port field is required. The value of the port field is almost always a number. On the endpoints of FDDI cables, the value depends on the type of FDDI port. If the port is a FDDI dual-attach port (i.e. it has an A and B socket), then the port field on the label will be A or B, possibly followed by a number. On all other FDDI endpoints, the port field will be a number.
For example, the following line might be found on a location label. In it, the Card and Port group specifies card A, Category 5 media, port 2. The rest of the label specifies equipment 21 in room 55of the Mesa Lab.
Site: 3 characters
Room: 6 characters
Sub-location: 6 characters
Connection Device Type: 1 characters
Connection Device Number: 2 characters
Card: 2 characters (one character for Card, one for "/")
Media Type: 3 characters (two character for Media Type, one for ":")
Port: 4 characters
Some networks devices are attached by cables that have one connector on one end and multiple connectors at the other end. These "fan-out" cables are represented by media types B3 and BS, as described in the previous section. The end of the cable that has a single connector carries signals for multiple ports. When labeling this type of cable, a range of port numbers is represented by two numbers separated by a period character.
Example lines with port ranges:
Many types of network equipment have multiple ports. For example, patch panels can have 48 ports or more. Creating a complete label for each port on such devices is wasteful, and is often not possible due to space limitations. To permit convenient labeling of such devices, a single label containing invariant information can be placed on the device, and shorter labels containing only the variant information can be attached to each port. The label that designates the whole device is called a generic label. An asterisk "*" character is used on generic labels to signal to a viewer that a field is displayed on a port label on the same device.
For example, a router might be labeled as follows:
Individual ports on the router contain labels that supply the last field of the label only. These labels would start with a dash character, indicating that the first part of the label is found elsewhere. This kind of label is called a truncated label.
For example, a port on the router from the previous example might be labeled as follows:
The guest network ports are active at all UCAR sites in some of the offices and in all Conference Rooms throughout UCAR/NCAR buildings. Each site is allocated a different guest subnet. The guest ports and cables are laballed using white print on blue ¾"width tape.
In offices, guest network ports are labeled minimally since users or guests often move offices or change subnets frequently. The label on the cable that is connected to the guest network port in the office will read as follows:
In conference rooms, guest network ports are dedicated as a permanent
port and are labeled more formally. For example:
To see the allocated Guest Network VLANS, see the
NETS Campus VLANs and Subnets page.
The maximum size of the labels are contingent on the physical size of the equipment.
Cards in equipment are labeled A-Z. These labels designate the card itself, not the slot. This allows a card to be moved from one slot to another without requiring that all the labels on the card and the cables connecting to the card be relabeled. A side effect is that the card labels found on a device may not be sequential, because cards may have been moved around after the initial card labels were applied.
The special Y character is use to handle ports that are part of the device itself. In other words, if a device has a 10baseT port for administrative control, and other 10baseT ports on cards, the special port is named as if it were on card Y.
Wall plates are connection devices, like patch panels. They usually have a generic label identifying the wall plate and truncated labels identifying each individual port. "B" lines indicate the location of the network device at the other end of the cable on the back of the wall plate.
In each room, wall plate numbers start at 1.
The maximum upper left label size is 3-1/2" long by 3/8" wide.
Each port label maximum is 1-1/2" long by 3/8" wide.
In October 1996, a new style of
wall plate was selected to replace 5-gang wall plates as the
UCAR-wide standard. The new plates are manufactured by SIEMON. The plates have a
swing-down door, which hides a slot containing a paper label
(see the figure below). These wall plates are
labelled with two identical generic labels, one on the outside
front of the wall plate and one on the top of the wall plate,
and truncated labels found behind the swing-down door.
Patch panels come in different sizes and shapes. Some have room for a label next to each port. The ones that don't have room have labels taped onto the front cover or elsewhere. These large, aggregate labels are kept in FrameMaker format. See Printing labels.
Fiber patch panels often have a single 19" panel containing two groups of connections. These panels are sometimes labeled as one patch panel, and sometimes as two patch panels. The choice is usually made based on where the cables on the back of the panel go. If all the cables on the back of a panel go to another patch panel, it's okay to treat the patch panel as one panel, and give it one number. If the two groups of connections on the panel go to two separate locations, the two groups must be labeled as two separate panels.
Note: port labels on the patch panel are not shown.
Telephone punch-down blocks, or "110 blocks", are found mounted on wall boards. There is limited space to label each connection, so on these labels there is no "L:" line. A large generic label to the left side of the block provides the local information.
Labels are usually printed using a DYMO 6000 LabelMaker. These LabelMakers are available at the front desks at the Mesa and Foothills Labs, and in the Mesa Lab and Foothills Lab computer rooms.
Some network devices, such as patch panels or Cisco 6509s, have no room for labels, even if the smallest sizes are used on a DYMO LabelMaker. For these devices, large labels containing many carefully spaced lines are printed on standard paper, cut to size, and taped onto some suitable surface on the device. To use them, edit a file, print it, trim with scissors, and attach to the device. To attach the labels, you can use scotch tape or double-sided tape to attach the printed paper, or you can photocopy the contents onto Avery sticky-labels.
The DYMO LabelMakers use 3 sizes of tape: 6mm, 12mm and 19mm (a.k.a. 1/4", 1/2" and 3/4"). It seems that 12mm and 19mm tape are sufficient for all needs, so we propose to drop 6mm tape from the inventories. In the future, we may decide to drop 19mm tape. The 19mm tape allows creation of 4-line labels, which can be useful in some circumstances.
Common sense should guide which size tape you use. The 19mm size should be used for all big labels, on racks, equipment, etc. The 12mm tape should be used for all cable labels. If you need a smaller tape, use scissors.
When you make a cable label, put about 11 spaces at the end of the text to make enough extra blank tape so that only blank tape wraps around the cable, not text. We don't want people to twist cables in order to read a label that has text that wraps around the cable.
If possible, when making the "other end" symbol on a cable label, use the arrow symbol instead of the less-than "<" symbol. It looks better, and is much more meaningful. Use the less-than symbol only when writing the label into a text file on a computer, where you don't have an actual arrow symbol.
Always use all capital letters on location labels and all lowercase on name labels .
Put labels on the front and the back of equipment, racks and patch panels. When you work on the network, you'll appreciate not having to walk around a rack to find out what something is. For equipment that has panel covers (FDDI concentrators, NSC routers) put labels on the front, back and under the cover(s), so you can still identify a device even if the panel cover is off. On patch panels with clear, removable Plexiglas doors, label the doors and label what's behind the doors, so if someone swaps doors by accident, confusion will be minimized.
There are six types of location labels: equipment or rack labels, equipment port labels, cable labels, connection device labels, connection device port labels and punch-block labels. Equipment labels label devices. Equipment port labels label ports on devices. Cable labels label ends of cables, and show the Local equipment nearest the label and what's at the other end (<). Connection device labels label patch panels and wall plates, and show the Local connection device and what's at the other end of the cable on the "back" of the port.The switch comment field
(Link to Visual Cheat Sheet)
Cisco Ethernet switches can store a short comment for each port. These comments do not affect the operation of the switch - we can store whatever we like in them. Cisco calls these fields "port names", but that can cause confusion, so we refer to them as "switch port labels". Switch port labels are displayed by "show port" commands, and are also found in the Port Lists.
Standard rules: when the switch port connects to
if the computer room has a tile coordinate system, the switch port label shows the tile coordinate and the machine name. For example, in the Mesa Lab computer room, switch port labels use a format like
The slash that starts the string indicates that it is using this new format, which NETS adopted in January 2005. There is no site or room field since these are found in the name of the switch itself. The characters between the slash and the asterisk are a tile coordinate, in this case "BG72". The characters between the asterisk and the colon are the name of the entity that is responsible for the machine, usually a UCAR division or group name, in this case "DSG". The characters following the colon are a machine name, in this case "bs1101". This standard fits as much useful information as possible into the port name field.
if the computer room does not have a tile coordinate system, the switch port label shows the rack, division and the machine name. For example, in the CG2 computer room, switch port labels use a format like
Here's a table of sample switch port labels, many of which are wrong. See if you agree.
|FL4-1334-M1-P2-C5:6||FL4-1334/M1-P2-C5:6||too many dashes|
|FL2-2077/West||FL2-2077-%-% West||use % if you must|
|CG2-3033-W1-3-138D||CG2-3033-W1-138D||too many dashes|
|tcom-gs-1 3/3||ok||link to another switch|
|CG1-0103-apc1-ups||cg1-0103-apc1-ups||use lowercase in names|
|CG2-MR-SNAP2.EOS||/#J3*EOS:snap2||machine name is lowercase with no dots|
|/J9*HAO:hinode.ds||/#J9*HAO:hinode||no dots in names|
|bad port||BAD PORT||use uppercase|
|CG2-2042/FA-R5||/R5*FANDA:%||"FA" is FANDA, R5 is "rack 5"|
|CG2-MR-ASTERIX.GLOBE||/#D4*GLOBE:asterix||no dots in machine names|
|GLOBE||/#%*GLOBE:%||use % if you must|
|CG2 T12-RAP CLUSTER5||/#T12*RAP:cluster5|
|CG2 Q12-cg2-mr-c1-ts||/#Q12*NETS:cg2-mr-c1-ts||NETS machine|
|was:cg2-roof-t1-wb||we don't care about 'was'|
|CG2 R14-APC1-UPS||cg2-mr-apc1-ups||NETS machine|
|NETS-VLAN8||SPARE||we never want VLAN info in switch port labels|
|FL4-1410-P2||FL4-1410-P2-%||use % if you must|
|maybebad||BAD PORT||'maybe' isn't useful|
|Bad||BAD PORT||use uppercase|
|Trailer||% trailer||needs more, like "NETS trailer" or "ML trailer"|
|/G4*NETS:netscm3.245||/G4*NETS:netscm3||no dots in names|
|FL2-2076/H2||/#H2*%:%||H2 is a tile coordinate|
|JEF-1A-W1A||JEF-1A-W1A-%||use % if you must|
|10Base2NetFeed||% 10Base2NetFeed||needs more information|
|cash-reg-sal-bar-21||% cash-reg-sal-bar-21||it goes to a wallplate, so use wallplate info|
|ML-34A SECURITY||ML-34A/R%-SASS:device||add more information|
"Private network cables" are cables that do not connect to NETS equipment at either end. For example, cables that connect storage components in the Mesa Lab computer room, or cables that connect wallplates in closets to support audio-visual equipment. When these cables are no longer in use, they are sometimes abandoned, perhaps under a computer room floor. When we encounter them later, they can be hard to identify. Neither end is connected to NETS equipment, so we don't know what is at the other end of the cable, who owns the cable, or anything else that might help us determine whether the cable can be removed.
To avoid these problems, NETS attaches a special label to both ends of any network cable that doesn't attach to NETS equipment. These labels are bright yellow, and are attached next to the standard location label, if one exists. These special labels have the following format:
HOST - HOST|
DIVISION/GROUP - OWNER - DATE
For example, a cable that connects a disk farm named DiskCluster1 to a cluster controller named CC3 might have a special label that looks like this:
DiskCluster1 - CC3|
CISL/OSD - Mark Genty - 2010-02-14
We try to follow this format as much as possible, but any useful information can appear on special labels.
I read the new UCAR Building Numbering and Interior Signage Standard. I think we can integrate it with our labeling standard. The new standard is focused on room names. It pretends to cover
"building numbers", but doesn't, so we only have to worry about room names.
Our standard gives a wallplate a name like ML-26A-W1-C2, where "26A" is the room designator. The simplest way to integrate the new standard with ours is to use the new room designators in place of our room designators. This may make for some long labels, but it can't be helped. Working from the drawings of rooms in the 3375 building, labels would be like:
FLA-3228-W1-C3 3rd floor, room 228, wallplate 1, port 3 FLA-2118MC-W1-C2 2nd floor, in "main corridor" 118, wallplate 1, port 2 FLA-2190MR/R12-P3-C1 2nd floor, in "machine room" 190, rack 12, patch panel 3, port 1
The building code could be something other than "FLA", like, "FL5". The building code for the NWSC could be "SC" for
"Supercomputer Center", or even (yuk) "NWSC". A label in a basement at the NWSC might be
SC-B156TC-P2-C33 basement, in "telecom closet" 156, patch panel 2, port 33
This approach would cause the least disruption to the NETS labeling = standard.