According to a ZDNet article, the average downtime of Windows Server 2003 was increased by 25% in 2007 whereas the downtime of major Linux distros such as Red Hat and Novell decreased by about 75%. The downtime of Windows Server 2003 is nine hours per server, RHEL average downtime is only 1.75 hours.
- Poll: How reliable are ChatGPT and Bing Chat? - Tue, May 23 2023
- Pip install Boto3 - Thu, Mar 24 2022
- Install Boto3 (AWS SDK for Python) in Visual Studio Code (VS Code) on Windows - Wed, Feb 23 2022
These are certainly interesting numbers. The article also explains why the downtime of Windows Server 2003 went up. It seems as if there were more Windows updates in 2007 requiring reboots. Hence, one cannot conclude that Windows Server 2003 systems are crashing more often than Linux systems.
However, downtime is downtime. These results match more or less with my own experience. Most of our servers don’t reboot, automatically. So I am quite aware of the downtimes of our Windows servers. I am even a little surprised that the average downtime was only increased by 25%. I’d say that the number of reboots was much higher than in 2006.
With every new operating system version, Microsoft promised to reduce the number of required reboots. Thus far, it seems to me that they were unable to keep this promise. Let’s see if it will be better with Windows Server 2008.
However, I am not sure if these results really speak for Linux. I think, the main reason why the number of reboots increased is because Microsoft is more serious about security than before. So the patches come faster and therefore reboots are more frequent. There have been quite a few studies that show that Microsoft is patching faster than Linux distros. Usually, Linux vendors reject those statistics. I never found their arguments convincing though.
One thing is for sure. The number of patches will go down with Windows Server 2008, simply because Microsoft’s developers value security more than ever. There is already a significant difference between Vista and Windows XP. The number of security updates in the first year after Vista’s release was almost twice as high for Windows XP. This does not only mean that Vista is more secure than XP, it also implies that the costs for patch management are higher with XP.
Subscribe to 4sysops newsletter!
Since Vista and Server 2008 share about 70% of their code, it is quite probable that we will see a similar development in the server field. I even expect that the difference between Windows Server 2003 and Server 2008 is bigger than with XP and Vista. Many of the teething troubles of Vista RTM were fixed in Vista SP1. Thus, the reliability of Server 2008 should benefit from these improvements, too.
WS08 already Benefits as being delayed release.
Nice tactic, ‘throw’ the workstation version in the trenches. Then apply the results to server rtm. I mean WS08sp1. whatever…
However, downtime is downtime.
I have to disagree with this statement. Downtime in the middle of the night, on a weekend, for a scheduled update, when no users are on the system, is a LOT different than downtime (e.g., a crash) during critical business hours.
I wish everyone would start being clear about the difference. I want to see “maintenance times” vs. “crash (down) times.” My downtime due to crashes over the last three years is more along the lines of zero (for most servers) to a few minutes per year and certainly no where near nine hours per year.
Greg, you’re right, that is certainly Microsoft’s strategy. However, that is quite okay with me. I rather prefer a problematic desktop OS over an unreliable server OS.
GlenH, the good thing about my sentence (Downtime is downtime) is that it is always true because it is a tautology. 😉 Seriously, I certainly agree with you. There usually is a big difference between a scheduled reboot and a system crash and not only because a crash might destroy data or the system configuration. I exaggerated a little in my article because I find those constant reboots quite annoying. However, it also depends on the environment. It might not be an issue if you reboot an internal file server during the night. It is something different, though, if you have to shutdown a web server that is supposed to be online 24 hours a day.