Thursday, March 25, 2010

Ubuntu Road Test (Final Report)

About two weeks ago I posted the FIRST PART (recommended read to get the full picture) of this article. The basic idea was to find out how Linux (Ubuntu 9.10 Karmic Koala to be specific) could cope with not having a recycle during a month period while being under intense use, regularly going from suspend to standard mode and back throughout the whole time.

Since that first post, not much has changed, really. I am amazed at how user experience has stayed consistent all along. Resuming activity from suspension feels exactly the same as it did the first time, about a month ago, and so do all other tasks.

In terms of responsiveness, clicking on the start button wakes up the system in about 1 second, making it responsive straight away. A fresh Windows box usually wakes up pretty quickly as well, but then it takes a couple seconds before the keyboard becomes responsive so the session can be unlocked. After that, Windows does particularly well resuming wireless connection, though. My experience (and this has been logged as a bug by many already) with Ubuntu is that resuming wireless connection from suspend mode can take quite a long time, sometimes even requiring user intervention in order for it to work. NOTE: I believe that this problem is down to the gnome-network-applet, though, as I am using Wicd on this testing machine and it resumes connection faster than Windows. Most times it has already resumed connection by the time I manage to unlock my session!! If you can live without 3G modem support, Wicd is THE network manager.

In terms of performance, even if I am almost exclusively using this testing machine for all activities, I still do boot my other Linux boxes every now and then. I wanted to keep an unbiased view here, so it was important to compare the performance from my testing machine to that of others which were not going through the same demands.

Once again, not much to report here. My testing machine just felt fresh and fast, and I could not tell any significant difference to how other machines would perform similar tasks.

As you may have noticed, I am not really timing application opening or suspend recovery times, for example, but that's because I really can't see any drag or slowness. Remember the point of my testing was not to measure response to the millisecond, but to find if performance would be as heavily degraded in Linux as it was in Windows PCs.

Surprisingly, I have not found ANY performance degradation on my Linux box. I am aware that perhaps the Windows machines that influenced this test did get a bit more "beating", but there simply is no contest when it comes to results. The degradation in those XP machines varied from significant to simply making the machine useless, but nothing that could even stand a comparison to my testing results.

NOTE: I want to clarify that the machines that inspired this test were using Windows XP. I am pretty sure Windows Vista would suffer from similar or worse performance degradation, but I must admit I am curious about how Windows7 would do. Please share your experience if you have tested it!


So there you have it, Linux passed this test with flying colors, performing consistently throughout the whole month under some intense use. In addition, I want to note that I have not had a single issue or crash during these past 30 days, which may not be a surprise under Linux standards, but significant when comparing it to other operating systems.

I always recommend people I know to use the software that best fits their needs. I am no die-hard Linux fanboi and have no problem acknowledging Linux flaws or weaknesses. Having said so, I still feel many people try Linux and simply follow their first impression. Eventually, it is mostly an exercise of "Well, this is not how I do it in Windows", and they just go back to what they know better. If they got past that getting-used-to phase, though, I believe Linux could add a lot of value in terms of performance, consistency, security and flexibility. At the end of the day, that all translates in higher productivity for the end user which, unless you are using your PC as a gaming console or a media center, is what it's all about, isn't it?

Thanks for reading!

No comments:

Post a Comment