Microsoft, it seems, is reaping the benefits of years spent patching Windows online, beating online update services from newer - and sexier - rivals. Windows Update enjoyed 100 per cent uptime during the second-quarter of this year, according to a survey by Pingdom. Windows Update piped Apple, which came second and trounced …
Uptime, according to Microsoft, is the time between mandatory reboots. But because you know a few seconds ahead of time that the system is going to be unavailable, they don't count it as "not up".
Which is, of course, utter bollocks.
I wonder if they're using the same lying scumbag redefinition of "uptime" in this latest glorious announcement?
While this may be true...
While this may be true, you have to admit that Ubuntu was a little less likely to be exploited if not patched instantly, unlike Windows. I spent a whole day this week cleaning viruses (virii?) off of compromised machines that may have been infected due to one of my colleagues not releasing Windows patches on our WSUS server, and/or users letting them languish without installing them. (and our Symantec "Corporate Edition" completely missed the infections at first) One of the three machines' XP OS was damaged enough that it still needed to be rebuilt after cleaning. While Ubuntu has its problems, some of which can be very frustrating, and harder to solve than Windows' hiccups, I found myself longing for my Ubuntu desktop machine at home by about lunchtime.
As a note, I accidentally got one of my thumb drives infected while cleaning up another infected machine a week or so ago. Symantec never noticed it, and after a long day, I didn't notice the 'extra' files either. (fortunately I keep autoruns disabled on ALL of my machines) I plugged it into one of my home XP machines and freeware Avast AV detected it instantly. Go Avast!
compelling reason for windows ?
I mean the fact that their update server is more stable than their product ?
So maybe they could use the knowledge gained from the update server to actually improve windows ?
I don't think the fact that the update server is very reliable is going to be a reason to buy their products. I would prefer a reliable product over a reliable fix-it-as-we-go-along product. XP Service Pack 3 anyone ?
"It should be noted, though, that Ubuntu's repositories has [sic] mirrors around the world, so users can download packages from those as well,"
So it would be 100% uptime from the client's point of view then? Even better, if you combined all mirrors, it could easily climb up to a theoretical 300-400% uptime surely. Which would leave Apple and MS far behind...
Need More Info Here
I'd be more interested to know why I might need the update service from these OS suppliers. For example, how many Security Critical updates were issued for Vista compared to Hardy Heron? Or even how many *still* for XP compared to Ubuntu overall?
While the article states that Canonical's update service was down, though mirrors were available to download packages; it isn't clear if the mirrors had update packages available. So, were Canonical's updates available from the various mirrors around the world even when the 'home' update service was down?
Further, the article starts off by comparing 'Windows' with 'Ubuntu', then later it refers to 'Linux' updates. Can we have confirmation that this word 'Linux' is meant to refer to 'Ubuntu'? Maybe I'm supposed to make 'reasonable assumptions' ?
For any of this information to be useful, we need to compare like with like and we need confirmation of what was actually meant by the terms used.
Well, any new updates released while Canonicals servers were down probably didn't get spread to the mirrors until the Canonical servers came back up, other then the few that might have been affected like that however, everything should be availble on several mirrors.
The odds of any major linux distro's distribution servers going down entirely are astronomical. Esp. since most of those servers run on *nix based systems themselves and tend to be a hell of a lot more stable then anything run on Windows.
Now if these figures are real uptime figures that also take into account any downtime by (mandatory) reboots, this would show that Windows can run for 3 months straight without major memory leaks. I'm guessing the update servers are running W2k or W2k3 server with an awfull lot of modifications.
Ofcourse it's also entirely possible that they looked at Windows update as a whole which is undoubtedly made up out of several machines, probably at several different locations. Now that would skew the results.
Actually the question is: Who Cares?
Who cares about the uptime of Windows Update Servers or Ubuntu...?
But even if one would be barely interested in these Update Servers, one might be more interested in *HOW* many critical patches have been addressed by the ... or better... *HOW* many critical flaws there are still that need to be patched. This one would be a comparison that I would really like to see.... but maybe impossible...
So, do we NEED Ubuntu update to be up all the time?
With Windows it's critical, not seen a really dangerous zero-day with Ubuntu yet.
But I may have missed it - was too busy with all the Windows updates:
Adobe (PDF, Flash), anti-virus, windows patches, various firefox plugins, oh, and using a printer over USB means that I had to re-install the %&%ç* thing from scratch as Windows couldn't recognise it already HAD the drivers (it started as USB, then lived on a print server).
It seems more and more of my computing environment appears to be designed to keep me FROM work or at least interrupt it as often as possible. Time to buy a Mac and avoid MS software, I guess. At least I'll get some work done.
Are we assuming the update server runs on Windows? This seems like good uptime for a Microsoft product, perhaps it's running 3.11...
Re: compelling reason for windows ?
>While the article states that Canonical's update service was down, though mirrors were available to download packages; it isn't clear if the mirrors had update packages available. So, were Canonical's updates available from the various mirrors around the world even when the 'home' update service was down?
That'd be hell to monitor, though you raise a good point. The fact that the Pingdom quote points out mirrors were available, without addressing the detail that if the primary repo goes down while or shortly after an update is made available on it, no other mirrors worldwide will have it until the primary repo comes back online, meaning theyre as good as "down" as well, seems to be purposefully lacking.
@wim - The update service that's more reliable than the product? You mean the windows server 2003 cluster (a microsoft product) running the update service, thats more reliable than microsoft products? News for you buddy, properly configured windows servers are every bit as stable as other platforms, and the whole reason clustering exists (on all platforms) is to further increase this.
pity about the client
Thanks to Windows update, the *client* computer can never have 99.9% uptime!
I love it when Windows decides to reboot overnight (I still haven't broken the Unix habit of leaving a bunch of programs running).
The update service running Windows came top..... Uptime rating it beat Apple (just) and Ubuntu....
Looking at it Wim, I'll think you find MS is running Windows for their own servers. It'll be Windows 2008.
Was that the server or the PC that was down?
Something to be remembered is that that vast majority of Ubuntu updates can be installed without disrupting your work. I had too many incidents when the beast of redmond forced an update on to my PC and then kept reminding me that I had to reboot now or reboot later. One moment away from the screen and the default reboot now takes over and just trashes any work you had open,
"Hello??? Microsoft??? It's my PC - you should NEVER, EVER assume you can take it over from me". I know I can turn this off, but it is never a problem I have had with Ubuntu (switched two years ago and I ain't ever going back).
Paris, because she knows how to handle uptime.
Microsofts WU domain is behind a load balancer on a content delivery network, and so it hides all the failed update servers. Which doesn't matter as long as enough are working.
Ubuntu doesn't hide them, it simply tries different ones. So passively checking ONE SINGLE SERVER is absolute crap. What you need to check is whether you can update or not, nothing else.
Spin, bollocks and crap, the whole article.
M$ 100% Windoze Update server uptime...
...and the patches only come once a month. If M$ deem the bug/vulnerability worthy of a fix. Maybe.
The points are well made about Ubuntu update servers being mirrored all around the world, but we also get updates much more promptly and much more frequently, we don't have to work on a system that remains vulnerable until "Patch Tuesday".
'Microsoft Update' that NEEDS to be available 27/7? I suspect they have lots of backup machines and bandwidth redundancy to make sure that this is the case.
Th fact that they NEED this I think is testament to the fact they, Microsoft, know how poor their software actually is. And when viewed in this light the article loses it's propaganda value.
What I find interesting is that while you can add extra repositories to the 'sources.list' text file, it is not done automatically. In such an event that the main server goes down then the mirrors can pick up the slack. Apt and Synaptic both rely on the'sources.list' on which servers to contact, maybe an alternative mirror list should also be included. Something for Debian and all it's progeny (Ubuntu etc) to think about.
Perhaps they're not quite telling the truth........
I found Windows updates inaccessible at least three times during that period and for more than an hour or two.
Now, that could be down to their server being "up" but taking too long to reply, but it amounts to the same thing as far as I'm concerned.
Don't use Ubuntu so I can't comment on that but at work I also use Centos and have never found a problem with that.
As an aside: Last week I changed my wireless router, which meant I had to point my PC to the new SSID. Suddenly Windows said I had to validate my license again!
The didnt use a new computer...
Recently a client of mine bought a new Dell laptop with XP just before the cut-off and when we got it, it was only XP Sp2 and majorly behind on updates. We ran the windows update and the first thing it wanted to do was download XP Service pack 3. We let it...
But after it installed SP3, no other Microsoft updates would install. We couldnt even get IE 7 to install. Every automatic update and install from network admin download versions failed.
After we uninstalled Sp3 we were able to install IE7. After IE 7 was installed we reinstalled SP3 and suddenly all the other automatic updates worked...
So if you have a new XP sp2 machine, and let it install SP3 first like it wants to, afterwards all of Microsofts updates are unavailable... Whats that in uptime? 5%?
You mean the patches server availability? So we're one step away from the reliable computers now, we're aiming for reliable patch delivering computers.
Cos our computers are so unreliable that they need patched and patched NOW!???
Ever think some blog missed the point?
Microsoft was up 100% of the time to enable fixes for their pitifully insecure software.
I guess Texas should claim an uptime record too; no execution in the last year was delayed because of the system being unavailable.
There are some "records" that just don't need to be bragged about.
As I understand it, this doesn't make any sense.
I'm fairly sure Ubuntu's update notification applet works just like most other distros'. It does all its work client-side. There *is* no update notification server. There's just a lot of Ubuntu mirrors, all around the world, as the article notes. The update notification applet simply checks if there are any updates available in the particular mirrors you have set up locally.
Given this, how can the update notification service itself possibly be described as having uptime or downtime? Unless all Ubuntu mirrors worldwide go up and down at the same time (which..er...they don't), that just doesn't work at all.
If I'm wrong about how Ubuntu's notification applet works do please correct me...but I don't think I am.
Microsoft Windows Update
sits on Ubuntu/Apache of course.
"Windows Update piped Apple..."
[Faux Scottish accent]
Aye! There's nowt like the skirl o' the pipes to put the fear o' God into the hearts of th' opposition!
[/Faux Scottish accent]
Is there still an urban myth to the effect that bagpipes are listed as weapons of terror, banned under the Geneva Conventions of War?
RE: Canonical's Mirrors.
Mirrors are exactly as the name suggests, a reflection of the original server. Hence, a mirror server should contain all the data of the main server otherwise, it's not a mirror.
multi billion dollar company has more server than $10 million foundation?
Wow, who would have thought that Microsoft would have more servers behind their update.microsoft.com domain than a foundation with a handful of full time employees and running on a few million in donated cash? Pretty impressive that such a small company gets compared to such a massive corporation such as Microsoft. But too bad it's just not big a deal.
For something meaningful, how about checking what OS survives updates better? I've only run across one issue with Ubuntu update messing up in 3 or 4 years and that was fixed like a day or two later.
It should also be noted that the Ubuntu update system updated thousands of applications besides just the OS. Microsoft updated only their OS and a handful of Microsoft apps and even then, they have a bad record at not blowing up the OS with the update.
PS: article helplessely biased
There a huge MS-centric bias in the article, while I'm at it. Closed proprietary systems require a constant support from the vendor's servers. Open source ones don't, so MS absolutely *needs* high uptimes to keep its world spinning, while open source OSes don't. Uptimes are therefore relevant only for MS, which is a *flaw*, not something they should be proud of. If there's a problem with my Linux boxes (not Ubuntu btw, why would I use that?), and say, I need a security patch, I'm pretty sure I can find the source code somewhere and compile it myself, even if the "mother" servers AND most of the mirrors are down. That's why I wouldn't worry about my Linux systems even if my distro's servers had a 50% uptime. But I would be VERY afraid for my Windows machines if MS servers had less than 99% uptime.
Oh, and in case you wanted to know, my distro's primary servers have a better uptime than MS servers anyway. Not to mention the mirrors.
What was MS bragging about, again?
M$ manages to keep something running for more than 55 seconds and now the article is crap. my my and no mention of apple being in the close #2 position. I think you guys need to lay off the caffeine a bit.
Will you lot ever stop moaning. At least a windows OS doesn't need updating daily.... I'm talking to you "crappy crud" or whatever you're calling yourself these days.
Microsoft updates 'faster' than ...
Having kept track of Microsoft 'updates' over the last few years...
and monitoring the ones that are years old, that have yet to be acknowledge...
and, that Microsoft has acknowledged that Active X (or whatever they may want to call it) can never be made secure...
It must have been the Redmond gang that paid to have that 'truly and honest'
survey taken and tallied.
Of COURSE!! I do not doubt its validity...
Just like I know the media reports all the facts, honestly... (yeah! right!)
Fair enough, if MSFT didn't have the update server up 24/7, we'd be deriding its instability. But this article is quite useless. How did they measure uptime? Responds to ping? Like Mr. O'Connor pointed out, uptime doesn't mean much if the updates don't actually, er, update.
Although, really, given Vista SP1 woes and WGA... isn't there times where we just wished the update servers were down?
"But after it installed SP3, no other Microsoft updates would install. We couldnt even get IE 7 to install. Every automatic update and install from network admin download versions failed."
That's because the update to windows update isn't included in the SP3 update. You have to update the updater before installing that update :)
I understand this is what caused it:
What a load of do!!ocks this so-called news. MS are continuing to lose the battle now, full stop (or period if you are American). They are staffed a bunch of useless MBAs who know SFA about the real world. Let them play with their crappy eye candy. MS were awesome until 2002 but then then the parasitic to$$ers won and the rest is history. Uptime!? Ha bloody ha! More irrelevant than my curried farts are to global warming.
Icon - curried fart exploding.
"At least a windows OS doesn't need updating daily"
ahahahaha MWARHARHARHARHAR *cough* *cough* cough*
That was a good one. Life must be easy in the strange galaxy you're living in... now you owe me a keyboard.
PS you'll find that in our world, there is a slight semantic difference between "is not updated daily" and "doesn't need updating daily". One can *theoretically* imagine, for example, that when updates as trivial as adding a couple words in a dic require a huge download and a reboot, our MS overlords *would* choose to wait and group the updates even it it means delaying important updates. We could also *theoretically* imagine that for people using a real OS, downloading a few kB twice a week to update the whole system *would* not be an issue. Especially as if it *wouldn't* require a reboot unless you'd be replacing the kernel. All this is theoretical of course. We all know that ANY update requires at least one reboot (if not 3), in the real world. It's not as if there were well-thought OSes around. *Theoretically* relevant icon joined.
Permit to use your own property
You don't have full control of your own computer.
Change your system hard disk (yeah, they're monitoring you) and you'll have to re-apply to renew your permit for your restricted software and hardware.
hang on. my 2 Ubuntu Boxes run an OS that is free. i dont pay for the OS and I dont pay for their wonderful update/patching service.
that said, although their main server was down, they have mirrors available everywhere and my system is set to use those. just like my CentOS boxes will use the fastest most local server (also free)
1 day of downtime - and never had a 'real' reason to update - my Ubuntu's are desktop
systems that dont run any external access services or network listening services unlike my windows XP and Vista boxes which i am constantly in a panic about (either for reasons or because the security industry is constantly making us paranoid!)
but still... packages.ubuntu.com shouldn't be down for 24 hours +
of course, it doesn't make sense that Microsoft is publishing this FUD. But I still don't understand why sites like packages.ubuntu.com are down for a day or longer, when one could also redirect requests to a mirror site. Also for people that are not in search for FUD arguments, lack of such a simple thing as redirection during downtime does not correspond to canonical's image as a reliable distributor of a very reliable OS.
Also, AFAIK, it is only possible to do a package search exclusively on the site packages.ubuntu.com, whereas this cannot be done on another mirror site. In that case, canonical's downtime _does_ count. And no, one cannot always use apt-cache search... for example when one just need to download certain debs for offline install on another PC. Rather specific, but still a significant use of packages.ubuntu.com...
Paris, because even she needs to do a search sometimes on packages.ubuntu.com, when she just needs the debs for an offline install.
Pierre explained this rather well and was very amusing at the same time, but perhaps his style was a little too subtle for a legacy OS luser like yourself.
(1) OS code is imperfect (like any bit of code, with the possible exception of Doom 3 - peace be upon it - which is a g(l)ory unto itself, amen) and that goes for MegaloCorp monopoly-capitalist M$ OS code as much as for free-range organic fairtrade hand-woven FOSS OS code like Ubuntu.
(2) When these imperfections are discovered they need to be fixed.
(3) Getting the fix once the new code is deemed stable is better than waiting until a particular day each month deemed astrologically most profitable by said MegaloCorp M$.
(4) Kiss my penguin's butt :p~
I'm fairly sure ubuntu uses a round robin configuration with their servers, where if one server is down then the next mirror in the ring picks up that address, etc. In which case it is more or less impossible for the update service to be unavailable as it would need every mirror (several in every major country) to be down at the same time.
"Windows Update enjoyed 100 per cent uptime during the second-quarter of this year, according to a survey by Pingdom. Windows Update piped Apple, which came second and trounced geektastic Ubuntu, which came a far-distant third."
Microsoft has tens of servers and several load balancing front ends, pinging those is irrelevant, it just tells that a front end is answering to ping, not that any of update servers is serving anything.
Totally meaningless number and claims done bollocs. Pingdom is obviously a subsidiary of MSTF.
Actually by default ubuntu uses your country-selected mirror (us.archive.ubuntu.com for example) to download normal updates, but security.ubuntu.com for security updates. I assume this is to speed deployment, but it also makes everyone dependent on one server for security updates.
I'd say their setup is a little more complicated than that, I'd suspect that they have GSLB in place, to direct you to the nearest (geographically) serverfarm.
Anything less than 99.999% these days is just not acceptable from a corproate entity.
Paris, cause she's had more downtime than Ubuntu's site..
Since billg once made the mistake of comparing the motor industry unfavourably to his, perhaps I could ask if he would prefer a cheap car that hardly ever needed fixing to an expensive one that broke down regularly, but had a good network of garages?
The client perspective is all that matters
I run Windows, Ubunta and OSX machines.
The OSX machines have given me no problems in the download department.
In the last three month I have had one instance where Ubuntu updates seemed to be offline and that was most likely because I had disabled mirror access and am still looking at "hardy-proposed". No problems though: I could keep working and the updates worked fine the next day. My other hardy machine (main hardy) has had no problems).
The friggin Windows machine auto rebooted my machine twice in that time. I'm pretty sure I'd told it not to reboot my machine during updates. Bloody annoying since I was once running overnight downloads and the other time I was running a long running a week long stress test that got killed after 3 days.
"Actually by default ubuntu uses your country-selected mirror (us.archive.ubuntu.com for example) to download normal updates, but security.ubuntu.com for security updates. I assume this is to speed deployment, but it also makes everyone dependent on one server for security updates."
Dunno about you, but every copy of Ubuntu server that I've installed has the default "archive.ubuntu.com" in sources.list - I know this because it (along with the AU mirror) never bloody works. AFAIK it's only the desktop versions that change the mirror to a localised one.
And all you knobheads claiming that Windows update restarts your PC after every update and thus can't have 100% update, RTFA - it's about the availability of update servers. And to those moaning about load balancing being the reason MS can have 100% uptime, so what? If Ubuntu (and thus Canonical) are better than MS in every way that you freetards keep saying, why aren't they able to put in a simple load balancing setup, yet the slow, stupid giant can?
by way of comparison...
Windows updated nicely this week, killing my PC's internet connection stone dead with two seperate security updates. I've heard of a couple of people suffering the same fate.
The laptop running Ubuntu updated with no problems whatsoever.
Thanks Micro$haft. You drive me further into Linux Land with every passing day. I owe you one.
@ AC and joe about *.ubuntu.org
"Lack of redirection -> people depend on a single server" and "24+ hrs dowtime in a row gives a bad image". You're both right. Canonical is not up to my standards either (and that's the *other* reason why I'm not using their distro -appart from Ubuntu distro being a bastardized Linux, which is itself a bastardized UNIX.) (1)
On the other hand, I definitely don't want a redirection when I'm dealing with security updates. That would be Bad Practice in all its splendor. And as stated before, the good thing with open source is that you can grab the source and compile it yourself, in case you need an important fix that is unavailable in binary form. You're not dependant on binaries. That's a tad less practical, but when security and stability are crucial, it can be a life-saver. Also, if you're really paranoid, you can (and should) review the code to make sure it does what it says on the box (if you're too lazy or lack the competence to do that, there are plenty of code wizzards who are actually doing it for you, which is not quite the case for closed, proprietary stuff).
Now a quick note for the dinos: I do know that x86 is what ruined real computing, but hey, it's cheap (especially since MS took over the world: you can get tens of perfectly good 10-yo machines -and parts- for free. You could almost be paid to take them if you wanted to.) and with a reasonable amount of redundancy a "free" (as in "I didn't pay for the hardware") x86-based cluster can hold itself against an expensive "big iron" piece of kit quite well. (yes, I've got some space. And I don't need no stinking heating during the winter: I just fire up my x86 cluster ;-) )
(1) DISCLAIMER: I do like mixes, and bastardized stuff.
Debian GNU-Linux is my choice for desktops and laptops (I'm waiting for a full GNU-Hurd port of Debian), bastardized allright but very practical (ready-made support for about any x86 hardware you can imagine, huge binary repositories, near-perfect package management).
DragonFly BSD is what keeps my cluster spinning (with a touch of genuine Open BSD here and there -hey, sometimes stability *does* matter) -this might eventually be migrated to Debian GNU-Hurd too, if and when it's stable. BSD lovers, flame on!
And I'm using a BlueBottle machine as my firewall (it's written in Oberon. Bring it on, script kiddies, try and find a kit to hack that! Dog bless cryptic OSes).
So I do like mixes, and "unlegitimate" OSes, but Ubuntu is just too bastardized for me. Namely, it has MSWindows-like genes in it. Barf.
Ubuntu doesn't automatically fall back to mirrors
Pierre: "So it would be 100% uptime from the client's point of view then?"
Sadly not. Ubuntu's update system doesn't automatically fall-back to the mirrors. I sense a feature request...
Windows update or crapdate?
I agree with AC its total bollox measuring uptime by ping response, my routers enjoyed the same availability, so what?
As for Windows update, Id much prefer that MS 'update' diddnt spew out bile that, in the past has caused perfectly good servers fail to boot & more recently create problems with Internet browsing with its recent KB951748 release.
Cheers Microsoft, another good reason to avoid your half baked slop.
Hrmmm, what is this test actually doing anyway? If it's just pinging the servers then that's not really a fair test. The application server (IIS / Apache in this instance I assume) could be offline, or the storage could be unavailable, or a million other things.
Pinging the server alone does not tell you service uptime - synthetic transactions need to be run against the service in question (something that I imagine is rather challenging with Windows Update).
I suspect all this was testing was the providers network connection and the VIP that their boxes sat behind...
- Mounties always get their man: Heartbleed 'hacker', 19, CUFFED
- Batten down the hatches, Ubuntu 14.04 LTS due in TWO DAYS
- Samsung Galaxy S5 fingerprint scanner hacked in just 4 DAYS
- Feast your PUNY eyes on highest resolution phone display EVER
- AMD demos 'Berlin' Opteron, world's first heterogeneous system architecture server chip