I think you should keep in mind certain principles in naming CIs:
1. It is much simpler if the name and the search code do not change throughout the life cycle of the CI.
2. It is much simpler if the search code reflects some information that is used on a regular basis, than using some esoteric value.
3. It is unlikely that the same convention can be used for all categories of CIs.
4. Depending on your version of OVSD, you may or may not be able to start a search code with a digit, which could radically change your strategy.
5. Like any other information system, a name should have more significant information at the beginning, which will make it easier for OVSD to auto-complete when you do data entry.
Some ideas: - for hardware, use an asset tag id for the search code. Advantage: doesn't change. Disadvantage: the granularity of CIs does not always correspond to the granularity of the asset register.
- for software, the devil is in the granularity - do you want to track products, versions, installed instances...? The answer will impact your naming convention. That being said, using the common industry name of the product as the basis of the search code is a common sense approach.
Pitfalls: - host names are likely to change. track them be all means, for example in the name field, but be careful about using them as search codes.
- DNS aliases are often the most common and useful names of networked CIs, but there are often more than one for a given CI, making it complicated - but not impossible - to use in Advanced Find.
- serial numbers often start with digits, so cannot be used for OVSD 4.5 as search codes. Workaround: prefix the serial no. with a very short abbreviation corresponding to the category. Disadvantage: more work to enter data.
Good ideas from Josh and here is some more food for thought.
When using OVSD it is nice if there is a few letters that group categories. Eg we have SRV- prefixes for logical servers so if when entering a CI in a Service Call or Change etc the user can just put in srv- and only servers show in the list.
We have one prefix for all applications (APP-) even though one might be a category of business system and other might be a desktop software. This is because the 1st level support may not know whether it is desktop off the shelve or a custom built business system.
We also have seperate CIs for Logical and Physical Servers. Physical is the Hardware and Logical is the Name in the OS and it has all the apps related etc. Logical is a Child of the Physical. This means when you roll-over hardware you just change the underlying hardware CI. Hardware Incidents are logged against the hardware so as the hardware gets re-used throughout its useful life, the hardware incidents stay related to the hardware.
In addition to the advice from the others, my main point would be to use a convention which your staff find obvious or intuitive; by that I mean, it is simple for them to guess what the name will be. For example, for physical items, use the name on the physical label.