Many years ago a system administrator came to ask me to support their business case to deploy a new time server for all the systems on the network to have synchronized time.

I read the business case, it was reasonable and well thought out; having good time server infrastructure enables proper management and monitoring of large (or any) networks. The request was for a dedicated time server appliance that would sync of a known time source along with a GPS backup (time stamps are encoded into GPS signals with an accuracy of a few billionths of a second); just a few thousand dollars in all.

The business case also listed the equipment it would replace. The current server was listed (which wasn’t appropriate for time serving for various reasons) and a Timex wristwatch. “What’s the Timex watch listed on the current equipment manifest?” I asked. The response came that “it’s Mike’s watch”. Perplexed, I asked why the watch was on the equipment list and learned that Mike came into the office every morning and checked the current time server for drift and adjusted it to match the time of his Timex watch; what time source Mike synced his watch against was unknown. As an aside, Mike liked to wear Hawaiian shirts so on learning this I quipped that the whole company ran on island time. If it weren’t for this business case I would have never known about the human component of our time servers or its overall poor design. At least we had a time server though.

Key takeaway: Always check the processes and design of the things in your environment.