Determining System Uptime, and A Bit More

I'm probably being way too picky here.

A new Community Solutions article in the Microsoft Knowledge Base entitled How to find Windows uptime? has been posted. The article is short (four steps and a summary for Option 1), but from my perspective this short article has a couple of errors, and / or at least could be a bit clearer.

While I know what is meant by uptime (how long the computer has been "up" or running), someone looking for this information may not know that's what to call it or what it might mean, so in the event they are able to find the article they may gloss over it because they don't recognize the term. Perhaps a brief description of what is meant by "uptime" in the context of this article would be appropriate.

Further, the steps detailed in the article (basically, instructions on how to run a command - net statistics server - and then parse the output) don't provide one with the information the article indicates will be provided. Granted, this is likely an academic detail, but the command really displays the amount of time that the Server service has been running - not the amount of time that the computer has been up. It is possible (but admittedly not all that likely) that the Server service has been restarted since the computer was booted.

As an alternative, I might suggest using Process Explorer. (Boy, it seems that I've been pushing PE a lot lately. Oh well. It's a great tool.) One can get a pretty good figure on the system uptime by examining the start time of the Session Manager process - smss.exe. Smss.exe is the first user-mode process created when the system boots. It is created by Ntoskrnl.exe and is a purely native application - it doesn't use Windows APIs. Instead, it uses what is known as the Windows Native API. An overview of the Native API can be found on the Sysinternals site at Inside the Native API. Gary Nebbett also wrote the book on the Native API - Windows NT/2000 Native API Reference.

No comments: