OSNews

The point is that using Ubuntu you have a 1.4% risk you have to wait longer than 10 full endless minutes to get your updates if you are only using two specific mirrors. That's very, very, very, very, very, very bad!!!! You can't use Linux because it's totally unreliable!!! Use only Windows, because Windows is allway's reliable and up-to-date and..... ehh, wait a moment here...

It sounds a bit strange to me they choose only one populair distro to measure the (implied) reliability of the complete Linux range. Canonical is big, but not that big. How about RedHat? Would they be a bit more reliable, just because they have more physical servers available? What about Linux from Novell (Opensuse, SUSE Linux Enterprise)? They have far more update servers available than Canonical. What about Debian? Etct. etc.

Do they really need something like this to put Windows in a better daylight? If that is the case Linux must be gaining more momentum than I was thinking. On the other hand - Microsoft is a company that wont accept any competition at all. They are willing to throw in billions of dollars amd lots of "payed research", just to stamp out something that is threathning to gain more than 0,000001% of their market share. Linux must be a total nightmare for them by now...

Yeah because Microsoft COULDN'T POSSIBLY make a good server OS that would stay up 100% of the time for 3 months?

Of course Microsoft can make a server that can stay running for three contiguous months. Server 2003 is a pretty good server OS. I'm sure 2008 is a good one as well.

The reality is that Ubuntu and Apple servers aren't down enough for anybody to really notice. It's truly a pointless article.

Now if Ubuntu or Apple servers were down 25% or more this article would have some relevance since the end user might actually notice the down time.

Netcraft "what's that site running" shows www.update.microsoft.com was rebooted one day ago. Average uptime is shown as 23days and max uptime of about 30 days

http://www.update.microsoft.com" rel="nofollow">http://toolbar.netcraft.com/site_report?url=http://www.update.micro...

Does that mean W2K3 server can only run 23 days before a reboot?

The point here is that the article is without basis and creates its own standard of measurement before declaring victory. There is no way anybody can repeat the benchmark to confirm its findings.

I'm sure Windows update runs on more than just one server, that was the report from the particular box (out of many) that netcraft got sent to.

Uptime measurement is pointless.

Having vast uptimes means one of two things.

1. You are an idiot that never updates.
2. You are running a cluster so you just migrate processes around and reboot servers as needed.

uptime scores = penis size

The point? I give you the point: this company is just being clever and getting a very popular promotion campaign for 0€. The recipe is simple:

-find a topic comparing products/systems with big communities easy to upset (in this case, OS, Windows/OSX/Linux)
-publish some data related to your activity and showing an advantage for one of the competitors, it does not matter which one, or if this study is reliable or not.
-wait for the flame war to begin and to spread on the internet in a couple of days.

That's it, Pingdom just used its own service to get a fantastic promotional campaign. They just pinged 3 servers during 3 months, and the result does not even matter, as long as it shows a difference between the competitors. Honestly, we should NOT EVEN talk about this, it's exactly what they are looking for. Pointless. Yet successfull, so congrats to them I guess!

Interesting - its funny that within a few minutes of your post and my post, our posts have been moderated down. Oxygen thieves from the said company moderated down people who think that their company is worthless.

The point is in my opinion that Ubuntu has more update servers worldwide. For me archive.nl.ubuntu.com might be faster and has a higher availabilty then update.microsoft.com .

How is that different from say... linux trolls around the globe doing the same thing? Or are you just upset because this article favors Microsoft and not linux?

Just because it's paid results, that don't rely on anything interesting. It's only free advertising (troll) on Microsoft reinventing the wheel and claiming their solution is better.

Are you really that naive to call PC World/IDC and Pingdom a group of Microsoft trolls? Where is the paid result? Where is the transaction? Just because it didn't say linux rules doesn't mean there is no truth to this. Now stop being an ass about it because it didn't put linux in the good light.

I second that emotion...

If only I didn't post a comment I would have modded this up.

The apearance of the comparison has some historical basis. Not so long ago, the mythical uptime to beat used to be Yahoo with it's BSDs showing uptime of a year or more. Being in University with the time to much around and reboot NT constantly, the idea of a year without rebooting was mindblowing. (I went the Linux based OS way, a friend went FreeBSD so we had fun jabbing each other about it.)

I'd need too understand how this comparison was monitored and analysed to have an opinion on it's value specifically. If results are from a balanced technological comparison then that's what it is. In the past, that was used by the other fan camps to indicate stability.

These days, the first thing across my mind is "hm.. no kernel updates during that year? Teeheehee... How do I get that auditing contract?"

But you see, server uptime is not involved here. This was an *availability* measurement, not an *uptime* measurement. Microsoft could provide 24 servers for Windows Update and reboot one every hour, if they so chose, so long as the incoming connections always hit a live one. Implying that this makes Windows a superior OS is mind-numbingly stupid.

The real headline here is "Microsoft engineers can set up a fault tolerant, redundant service." Frankly, anyone can do that. It takes talent and planning, sure, but anyone with virtually any platform can make that happen. It says a lot more about the people who set it up and maintain it than it does about the quality of what they maintain.

There was another comment after my post that pointed out the Windows Update cluster uses a Linux based OS back end also. Based on availability and actual platform used in the cluster; I'd say it's good news anyhow.

So, to provide a valid comparison, we'd have to measure uptime and availability against single server systems.

Windows Live servers are different servers than the Update ones. Maybe that has something to do with it.

prove me wrong!

Get yourself a local repository.

I do a lot of building VMs for various whims. Besides the time it takes for the updates to transfer and install, there is also the consumption of my monthly bandwidth alotment (ah, big name ISPs). My solution was to write a quick script against rsync and mirror the repositories on my local NAS through http. I then leave them open to my lan through my own httpd. Another script cleans out and adds the local NAS repositories. A quick rerun of the repository sync script captures any new packages. If the repositories are slow then cron the script. The rest of the traffic between various builds installing or updating happens inside my own network space at full speed. The only slowdown is the actual installing step.

With Mandriva 2009 out soon, I'll have to see about adding in another mirror set. Hopefully I can confirm that nothing is still running 2007.1 and drop that mirror out of the NAS to free up space.

Burst into flames an disappear? I'd be cheering at that sight!

Stop being Microsoft.

a loose, loose situations? Sounds kinky.


"lose".

How about "Do nothing special, get article praising your product"?

How about we change the headline to "Windows server farm run by smart Microsoft engineers provides better availability than a single Linux server"? How about we call it "Better planning allows Windows Update to be available 1% more often than Ubuntu Update"? How about next time somebody tries comparing apples to apples.

This 'article' doesn't prove one damn thing about the quality of any of the operating systems involved.

They are reporting uptime via sampling (periodically checking on the server to see if it is up).

In this particular case, they are reporting on Microsoft's Windows Update servers, which is a cluster of Linux systems.

The article is silly because it imputes that Windows is more stable than OS X or Ubuntu Linux on that basis. That's a foolish conclusion for two reasons: they weren't looking at any Windows servers, and, second, a high-availability cluster is always (well hopefully) have 100% uptime (assuming it was setup properly).

And of course, both the Windows and Linux boxes are on the exact same hardware and running the exact same apps coded by the exact same developers, right?

akamai doesn't host it. They use them for caching and what not, like LOTS of companies do.

Actually they don't host the pages but they mirror the content. That includes a lot of the actual patches that you install. Also the images etc you see on almost all MS pages for almost all their sites.

If you are installing security updates to reboot every week then you are doing something wrong. Microsoft sends out patches once a month on the second Tuesday of every month. I'm not sure how you can be prompted weekly when the rest of the world does it only once a month.

The guy did make a valid point. We are here to discuss technology, and most of the spelling mistakes made by people here, are made by die-hard Windows users. Users who happen to use IE.

The other guy made the point that most of the spelling mistakes can be eradicated by installing and utilising firefox and firefox's spelling addons.

Therefore the original post is technology related.

I disagree, maybe some posters are using work computers and it is against policy to install firefox. Maybe they're using OS X (I do not know if it has universal spell check or not, or if safari does, I never checked). Maybe they are not native english speakers. Maybe they are canadian like me and spell colour and honour with an "u" (Firefox marks that as an error by default).

I could even get with the sentiment about spelling, but being nit-picky about grammar is just asinine, as a lot of people here are non english speakers, and firefox doesn't help that.

They do have localised dictionaries, hence the UK spelling of "localised"

"Therefore the original post is technology related."

Too bad it has absolutely no relevance to the topic at hand. Maybe we can have some posts about who makes the best microwave oven as well, since that's "technology related" too.

Please, no! The semi-literacy around here is my main source of humour - from the "funniest errors" text file on my deskto; (all courtesy of OSNews comments):

"for all intensive purposes" = "for all intents and purposes"
"died in the whool" = "dyed in the wool"
"server allergies" = "severe allergies"
"premisquess " = promiscuous
"Common poeple" = "come on, people, ..."
"provail" = prevail
"technicien" = technician
"pooring" = pouring
"heatred" = hatred
"holy grale" = holy grail
"flexibel" = flexible
"reprocutions" = reprecussions
"blote" = bloat
"plunty" = plenty
"infrigmentation" = infringement

Really? You really keep a textfile like that?

O RLY? Ya rly. Srsly.

What only 109% ?? And it reboots only 1 times per century! Great