Discussion Locked

PC Naming Standards ... what are yours?

0 helpful
45 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

I am curious as to what organizations are using for PC naming standards, and if there are documented standards anywhere. I recently got into a discussion on the topic with someone who insisted that naming PC's with user names is the best method. Our current environment includes over 500 nodes at 18 locations throughout the US and one in Asia. I strongly disagreed with the idea of using user names for a number of reasons, primarily because we had used it before and it was very difficult to manage. We currently use the following convention: (department or location and PC number). Here are some examples: ADMIN01 for a PC in the Administration Department
NYCSR01 for a PC in the New York City Showroom.

I am doing a study of what others are doing and if I could find documented standards that would be great!

Thanks!

45 replies

by notsochiguy

My employer has about 30 different office locations. Each one is given a 2 digit code (01-30). Those are the first 2 characters from our computer names.

The second two characters are derived from the department. For example, Human Resources is HR, Finance is FI, etc.

Finally, the next 8 characters are derived from the beginning of the PC's serial number; giving us a 12 character PC name.

An example would be 06HRHZ71Y300. If a ticket came in with this as the computer name, the help desk automatically knows which facility to which the call should get routed, and the specific department.

We previously used usernames, but the issues with that were two-fold: multiple-users using the same PC (we are in healthcare, so we are open 24/7/365) and turnover (I'd hate to be user John Smith, and my Pc is named Jack_White...after the guy I just replaced).

I'd say the naming convention you are using now is fine; especially if it is working for you...which is always the bottom line.

by dafe2

Actually you gave me some ideas....we (as well) never give a PC the users name.

We've allways named machines relative to the position & include the asset number.

Considering our business structure seems similar, I think I'll give some thought to your naming convention. I like that it's more granular to what we're currently using. Thanks!

by pr0teus

Just don't make them the names of the users! :D

Love the concept of using comp names based on facility/dept et al.

We use (for one of our clients - a small company with 8 machines) admin01, admin02, secretary, editorial01 etc. They use that same name as their usernames - which they don't have to stress about remembering since it is returned to the login prompt by default each time you turn on a machine. Not to say that they're NOT to ensure that they remember!

But we don't use an individual's name, mostly because of employee turnover. And we've found that initially (the previous IT service provider used employee names) there was a sense of ownership when your computer has your name on it, thus a willingness to download or load any crap onto the machine! Now, the names give a subtle reminder that these are work comps.

by mbradley

I totally agree with you. We use the CompanynamePCXXXX for desktops and CompanynameLTXXXX for laptops. To use personal names or character names based on scifi movies as I have seen before is childish. And, to one of the replies...to hell with a sense of personal ownership of the machine by using user names. Login names follow the ILastname scheme. If they cannot remember their first initial and last name, perhaps they need to look elsewhere for employment.

These are company assets and property, not personal toys. We track the machines for financial reasons by asset tag and services provided by the extension number of the phone number the machine is located at. This all works very well and the company is more then pleased by the professional manner in the scheme that I have chosen to employ based on past good practice and experience.

by netgeek84

I really like the way you named your computers. The way I have named out computers is to use an abreviation on the department and 3 digit numeric identifier and the month and year the machine was last rebuilt. This works well for me because if a new employee is comming in and I see that the machine hasn't been rebuilt in 6 months I will take the machine flaten it and rebuild it before the new employee comes in.

by tim.doyle3

I work in an office where technology is... well... rather 90s. We have three computers on 98, two on Windows 2000, two and a half on XP (my personal laptop comes in to play once in a while), etc. We are in the early stages of building a network in our new building, getting a file server/network fax machine, etc... as I configure each machine, I'm renaming it using the following

- MCP - the name of our company
- Operating Systems - 9x, 2k, XP
- Location - 001 is the owner's PC, 002 is her husband and so on around the office
- Department - SA for Sales, IT for IT, AD for Administration

In otherwords, my station would be called MCP9x003IT (even though I fit under three deparments, I'll claim IT as mine). That's how I'm planning to go about it.

Hope that helps. This thread has been very interesting for me while I redo our network!

by apotheon

I quite like the philosophy of your naming scheme, and it's much what I'd recommend for an organization that has to track large numbers of end-user systems. The only problem I have with your specific implementation is that the serial number segment of the computer name is too long. I'd definitely recommend using some shorter character string there. If unique identifiers can be derived from the serial number that are at most four characters long (preferably shorter), that would I think be the way to go. If not, an arbitrary but consistent system for producing two to four character strings would strike me was the solution.

Of course, for the network at my consultancy job, I'm the only person that really needs to keep track of most of the system hostnames, and it's a small network, so I can get away with pretty much anything I like for a naming scheme. As such, I've taken to using the names of mythological entities. They're just easier to remember for general purposes here. Go fig'.

by caestelle

every peice of hardware we have, has a corporate tag number and that is the name that we give to the PC. We have three different offices and we use this same naming convention in all three.

At a smaller company that I do some work for they use state names for PCs and country names for servers.

by it_lobo

In my equipment inventory I assign an ID number. I use WS for work station and assign a number after it. Example: WS-015, WS-034, WS-045 and tag the PC with the ID (All but the servers have this naming scheme). In my company we have only about 50 PCs and we have been upgrading and hiring alot of people this past year, so people are getting assigned new PC all the time. I like using this method because I never have to rename the PCs to someones name (that is what was done before).

As long as you document which PC is for which user you will have no problem. For the most part none of the users need to know the naming scheme.

by stress junkie

I generally like the naming scheme described by saigman. Any computer naming scheme should also incorporate a mechanism to obscure the purpose and location of the computer to people who are not involved with support while making computer identification easy for the support people. The scheme described by saigman begins to address this issue but then fails when it identifies the business department that uses the computer.

Generally I would say that any computer name that includes the name of the user or the department that uses the machine is a security problem. This is particularly true with M$ software. Anyone browsing the network can see the names of employees, which computer they use, and which computers are associated with each department. This makes identifying targets of malicious activity very easy.

In saigman's name scheme the first part of the name is a code for the office where the computer is located. That's good. The code is meaningful to IT support personnel but its meaning is not immediately clear to anyone else.

The second part of saigman's name scheme is a problem because it clearly identifies the department using the computer to anyone. It does this by using generally meaningful codes to identify each department. If the scheme was changed to have a numeric code for each department then the meaning would not be immediately clear to network neighborhood browsers.

Then, within each department the desktop computers could just be numbered. So, overall a desktop computer name could typically be something like 010202. Even though the long term employees might figure this naming scheme out it is still valuable since it obscures the pc function, owner, etc. from guest computer users or anyone else who has no business knowing the names of network resources and nodes.

This also applies to servers. I hate to see a server named 'payroll' or 'hr'. Talk about putting a target out in plain view!

So I believe that computer names should be meaningful to IT support personnel but should ALSO be meaningless to everyone else. This way even a computer guru with bad intent would spend a lot of time trying to find an appropriate target. It only makes sense to do whatever you can to slow the bad guys down. The longer it takes them to find a target the longer it will take for them to perform their mischief.

And it's an employee privacy issue. The fact that John Doe works at your company is private information. Employee names should not be made easy to harvest. Employee names should not be used for naming network resources. Also, I am totally against businesses that publish the email addresses of their employees unless those email addresses are just the names of departments or functions, such as support@xyz.com or sales@xyz.com. Making employee names available to a everyone on the network or, in the worst case, everyone on the Internet, is a violation of the employees' privacy.

by notsochiguy

I really didn't give thought to mentioning any security issues...but you are 100% right.

I think the committee that voted on the standard (I am not on the committee, so I can neither take credit nor blame...this all happened before my arrival) didn't really think this was cause for concern, since our users are so locked down, they cannot browse the network (nor can they install apps without IS doing it, or launch a command line).

Of course, a lot of companies don't lock down the users, so this standard wouldn't be the best; and there is always a way around security (or will eventually be), so even locking down isn't a guarantee.

Like I said, very good points!

by tomsal

We use a "WS###" name convention as well, preceded by the name of our company (which is a four letter acronym...but no..not the one you just thought of..lol).

Anyway in the network description we do put the users name in there. So in net neighborhood you'd see the the NETBIOS with the naming convention but next to it you'd see the user's name.

With the frequency users are moved around in this place and the fact that we must do updates manually still -- it would be more of a hassle if we didn't include user names. Just another table or chart we'd have to look up against the netbios name to figure who it is. So yeah I guess you have a point with security, but there comes a time where you have give up some security in favor of saving time and frustration on a daily basis.

Furthermore, we do have strong passwords on user accounts, we do have firewalls running, we do have systems locked down with user policies, etc.

So its not like there is zero security.

Finally, we are big into logging and monitoring -- we keep a lot of logs on what goes on. From end stations to servers.

by stress junkie

Security always involves drawing a line in the sand regarding what we are willing to do. Let's take the example of our own homes. Most people implement a level of security that makes it difficult for someone to break in without doing any damage to the house. On the other hand few houses could withstand the determined efforts of someone wielding a sledge hammer. It is impossible to completely secure any resource. We are left with deciding what level of risk we are comfortable accepting versus the level of difficulty or expense required to increase security.

by choppit

You've made some excellent points I hadn't considered. I'll bear these in mind for next time.
Cheers

by tonythetiger

I agree about the emloyee names (Kind of a PITA to be changing them all the time anyway). I'm not sure I understand about the security. Firewalls and other security (user security) should prevent those kinds of problems. The worst that can happen is someone can attach to a shared printer. There is no sensitive information stored on the PC.

by choppit

The point stress junkie was making about security is that if a rogue LAN user wants to obtain accounts information for example, a good place to start would be to find the names of the accounts staff. With this, (presumably easy to find) information it's easy to create a list of target hosts. Regardless of how secure the hosts themselves are, this extra measure helps to 'camouflage' the target.

by apotheon

If your network is in any way at all network accessible to the outside world, stuff like firewalls will not provide a guarantee of security. Obscuring intent in computer hostnames is a very good practice to adopt, in part because if someone cracks security on your network and this is noticed, the additional time spent trying to identify targets will give network admins time to try to log and stop the malicious security cracker's activities.

To assume that just because you're using the most-often recommended security measures you're "secure" is hopelessly naive. These security measures don't make your network impenetrable; they only make it incrementally more difficult to penetrate the network. "More secure" doesn't equate to "totally secure".

by pr0teus

Y'know, enlightenment is a heck of a thing!

Forgive me if you happened to read my previous post, I didn't read yours until afterwards. As usual, as IT professionals, sometimes we overlook issues such as those you just raised when frankly they are fundamental (or at worst case we need to start looking at things this way since obviously crackers are getting more organised, deceitful and determined).

VERY salient points (however, I don't necessarily agree that an employee's "company" email address is their private property - hence no violation of privacy. It's there for biz purposes)

by stress junkie

I was just saying that lists of employees are confidential. I agree with you that email contents are a business asset, not a private email account.

by jaqui

a department id that is a number. ( chart stored in a locked down file on one machine )
followed by ip number, for static ip.
if using dhcp, I add a grouping number instead of the ip number.

all id is based on a hostid, that is assigned to a specific mac address.

each system has a hostid created on it. this id uses several specific hw addresses in it's creation, and no 2 systems have ever come with the same hostid.
( lm_hostid from flex license manager )

by afram

I used to designate desktops as WORKSTATION# and laptops as LAPTOP#.

Servers were named after mythological places or personalities that somehow related to the service.

For example, the mail server was named after the Roman god, MERCURY, who was the messenger of the gods. The storage/backup server was called ALEXANDRIA after the Greek library.

by avuncular

Our international company uses 2 letters for the country, 3 letters for the office location, 9 letters for the machine name, which is the first 9 letters of the user's domain username, and 1 number in the end to allow for up to 9 workstations to be assigned to a user (we don't use zero). E.g., Jose Rodriguez from Portland, Maine office in the US has a machine name of USPRTJORODRIGU1
Believe me, it works well.

As far as the servers go, they are named with the first 5 letters as above, and a 4-digit #:
USPRT0001

by choppit

For servers I use a descriptive name (followed by a number if necessary).

For desktop/laptop systems I use a 2 letter prefix to denote the company, followed by 2 letters to denote the system type ("PC" for desktop, "LT" for laptop), followed by a 3 letter numberic index.

by ippirate

In my organization I use an ID + numeric coding.

XXXXXXXXabbcc where

XXXXXXXX = Division name
a = Building
bb = Department
cc = Unit #

The organization is still small enough here that I don't have to many variables to deal with. Two divisions, two buildings, x departments and 50 nodes.

I chose the convention because it provides reasonable understanding of location and peculiarities when working remotely but it also doesn't directly target any one machine by association/name.

by killerb

due to the fact that my company has branches throughout Europe, Americas, asia and Pacific, that is all part of the same AD forest. Which in turn has trusts between its parent company and all the other parents' companies subsidaries throughout the world.
We use a combination of country initals, company intials, whether PC or LT, date purchased (for ease if any issues under warranty period, can tell immediately) and finally the cumulative count of what number that machine is.
i.e. NZCHHPC16080585
works for us

by thecaveman

I work for a community college district with 12 sites. Our naming conventions cover almost every piece of equipment on our network. Our names follow a pattern like this: RPS86-RT2 (not a real device).

The first letter is the campus, in the example, R for Riverside. The next two or three letters are the building, in the example, PS for Physical Sciences. The next numbers are a room number in the building, for example, room 86. The next group is the type of device, PC for desktop computer, RT for router, SW for switch, etc. Finally, the number at the end is just sequentially assigned, in the example, 2 for the second router.

by bfilmfan

There are several different ones for the current client. They use WKS and then last 5 digits of the serial number for workstations.

For servers, they use a three letter location code followed by the floor and then a code to name the OS, followed by a code for what the server is and finally a code to show the instance of that server (like Exchange server 1 is 01).

Once you've learned the code, it is pretty easy to figure out if a server is virtual or physical, what OS is on it and what it does.

by ole88

We use an approach that makes it easier to identify machines that are due for a refresh (replacement). The first two characters are the division of the organization, then a dash, followed by a two character designation of the type of machine and then a three digit number. The first number represents the last digit of the year or purchase and the next two are a incremental number. As an example, the 10th computer rolled out that was purchased in 2003 from Dell and is a standard workstation for the purchasing department would be named PD-DS310. This way we know where it is, how old it is and what type of machine it is.

by tonythetiger

We have about 3500 PCs and laptops, and about 1000 other devices (Servers, routers, hubs, switches, etc.)

AAAbbbcc##d

where:

AAA Major location (Region)

bbb Minor location (office or County)

cc OA CD DU HS, or LT (office automation (standard machine) or cadd, Dialup remote, high speed remote, or laptop)

## Accounting location (inventory control)

d unique ID within above ##

Our servers, routers, hubs, Wireless access points, etc. Use a similar scheme except linited to 8 characters

bbbCC###

where:

CC FS,PS,HU,AP,RT, etc.

### Unique number

by jdclyde

We currently used first name, last initial. Not the best solution in the world, but at least it means nothing to an outsider.

servers turned into cartoon characters.

Started when I got a test server that was big for it's purpose so I called it huey, for baby huey.

When the next server came in, my coworkers thought it was part of Donalds nephues, so the named the next one Louie.

We went of course to Duey, donald and Daisy.

Our only windows box of course had to be named Goofy. If was too ironic of a fit not to!

by t-cally

We assign the last three digits of the pc ip and the location that it is in. I'm in a three story building name the POC, so an example would be POC(section)-floor#-122, POCadmin-1-122. The IP stays with the location not the pc. This makes it easy to track your IP and to no if someone has illegally grabbed and IP. If there is a breakout your location time is minimize because you know where the pc is. I have a print out of the floor with the ip listed in the coordinating cubical. (Windows environment)

by themessenger

Our workstations use a three part naming scheme.
3-letter company prefix, 3-letter pc function, and the dell asset tag
Because we use the same image and across the country, we have a script that auto-names the computer during Sysprep using the above scheme.
We have two company names, so based on IP, the script can identify the company prefix. Based on three specific workstation functions, it names the middle, and then the suffix is pulled from the BIOS through WMI. We then have scripts in place from the network management side that do the rest.
The Dell asset allows us to use the dell site to track all asset info.
The build type allows us to target specific software to the computer after build.
The company prefix allows it to join the correct domain and OU.
--- Gotta love automation.

Out servers are a bit different.
aaabbbb00, where aaa is location in country, bbb is function of server (app for application, nfs for file server, prt for print server, etc..) and then a 2-digit number.

This probably doesn't scale for large enterprises, but we are a small to medium so it works well for now.

by pr0teus

any chance you could send me a copy??? Stripped of any sensitive info of course (or replaced by dummy info)

by guru of d0s

Every year, they manage to find names for hurricanes and tropical storms by using alternate male and female names in alphabetical order.

We manufacture a range of computer based simulator systems, and each computer has a name so they can VPN to each other and to our office and be uniquely identified by name so we know which computer is which when they connect.

Using the Hurricane Naming Convention, the first batch of six systems were Alice, Bertie, Clara, Donald, Ethel and Freddie. We're still going strong using this system and using a book of names (intended to give parents ideas for naming a baby), we should be fine right up until 2054 or thereabouts...by which time I will have long retired and won't care anymore!!

by digitalxeron

I personally give each of my machines, since I run a server farm, their own name... for example, the main DNS and Web server is caled "The Core" while the backup data server is called "The DataHold", for mass amount of PC nodes, I'd use the format
PC-[A]-[B ]
PC - designates a PC
A: location, 3 char
B: last 2 octets of the IP with a dash replacing the dot

for example, your "ADMIN01" would be (ip was 192.168.22.01 for the sake of example)
PC-ADM-22-01

by theidemann

When it comes to PC naming policies, it can be kept quite simple depending on company size. Stay away from using the persons name (first or last) on the PC. What you have now is fine because it shows the region and location of the user. It is helpfull, however, to add department information into the PC name, or at least an abbreviation (i.e. ASIA-TOKYO-HR-05, NYC-SR-ACCT-01, or NYC-admin-01). This will make your job easier when it comes down to budgeting, and you have department managers coming to you and asking how many PC's they have to have updated, and what it's going to cost them. Just make sure that when you create your documentation, that you add all necessary information in case you need to troubleshoot that particular PC. It's also a good idea to have some sort of lable or ID visible to the user. If the user using NYCSR01 calls you, they should be able to call you already knowing their PC name, you pull up the PC names document, and then you know which department their in, the floor that their on, PC make and model, etc.

Best wishes,

Timothy.

by mr. tinker

When I was put in charge of one company's extranet, (40 nodes) I set about rebuilding 1 by 1 all systems and bringing the infrastructure up to date by "rollout". This included naming conventions. Being an "Ancient Peoples" buff, I used a mythology schema and stuck with it naming machines like "Argus-Mythological beast with 1000 eyes" the warning center console that watched over all the other systems in the network, the 2 dogs that watch over the gates of hades. The idea, that even though there is still much debate over naming conventions being a "security hazard" from all that has been seen to date it doesn't matter all that much what you name it, it's how you protect it. If its got a name-its a target no matter the name. The naming conventions I put into motion gave me and my staff a very clear connection mentally what was where and its purpose. And even made for interesting mettings, encouraging my staff to be a bit more creative, since they were taking care of mythological pets! It became a little more personal for all involved. And my superiors were thrilled with that kind of involvement by all of us. As in: who is coming in for the weekend
"Feeding" (maintenance and failover testing) of Cerberus and Kerberos.
A bit of whimsy and strategy combined means when your discussing strategy over lunch others overhearing your conversation won't easily understand your targets. So a breach from within is less likely then the NYC-R(outer?)01 or the ADMIN(istration pc) station. Sometimes, its all in the symantics.

by tomsal

Good comparison. Actually, with the exception being if you live in an older home (as my dad likes to say..."They don't build them like they used to") where it was brick and mortar quite literally, our homes are really pretty easy to smash through.

A sledge, a hammer, a bullet...well pierce through the doors like a hot knife through butter.

by chriswales

We use the assigned Fixed Asset number as the computer name as this allows us to track by inventory and user.

by himatech

I think naming the PCs like the AS/400 naming conventions is better, like:

work --> wrk
Order Taker --> ordrtkr
Manager --> mngr

by tony velarde

We're a mid-tier single site Law firm and we utilise phone Extensions i.e. EXT308. We have found this to be useful when users call for support. Our phone displays the Extension the user is calling from and we can jump straight on to their PC's using VNC Remote desktop.

by kdodd61

Use the property tag number followed by a dash and the o/s version.

009999-WXP XP
009999-W2K Win2k
009999-WXPL Laptop

Database contains inventory record of each device:
-Hardware, individual s/n's of each part
-Installed software
-Admin password
-MAC
-Network location-which switch, jack number, port number, user name, etc.

by banzilla

We are an all dell shop (Good or bad)anyway, I use the following format
model number-user's initals-service tag
so my machine is 4300-ss-2g9bv11. We are a small 50 machine shop one location.

Sam

by woodycomp

I've worked a couple of places and for a business I have always liked this naming standard.
O.S. 2 letters
L/W/S workstation laptop or server
1 letter for company or site
3 letters for Department/group
Numbers starting with 100A or 10A depending on size of company.
e.g. for XP workstation at IBM, finance department.
XPWIFIN100A
The reasons are that If I need to look in AD for specific machines I can find all XP, 2000 etc Workstations or Laptops for any department rather quickly. I support two sites also so that is why the company name is listed. The numbers are just to have an idea of how many machines are out there as I'm doing it. I never liked naming them after the users because what if the move etc. My thoughts are it belongs to the group who has the PC and the associated software is installed for that group. Hope it helps.

This discussion is closed to new replies.