NTBUGTRAQ Bashes Windows Update 565
BigBadBri writes "Russ Cooper, keeper of the NTBUGTRAQ list, has a few concerns (to put it mildly) with the trustworthiness of Microsoft's Windows Update."
This discussion has been archived.
No new comments can be posted.
NTBUGTRAQ Bashes Windows Update
Related Links Top of the: day, week, month.
Ya'll hear about the geometer who went to the beach to catch some rays and became a tangent ?
I don't trust Microsoft... (Score:0, Insightful)
So? (Score:5, Insightful)
Oh, by the way, youre car is just fine. No, no recalls at all for it. Well, one, but it's only important if you actually drive, so you're fine, I'm sure...
it's better than nothing (Score:5, Insightful)
Re:Trust? (Score:4, Insightful)
How soon until they don't tell you that and just start reporting your web browsing favorites and selling that information to others?
I like Windows Update (Score:5, Insightful)
Sure, like any given piece of software, you may run into glitches and bugs at some point. But, overall, Windows Update has provided me with an extremely easy and painless way to keep my systems updated.
Even my Mom can use it, which says a lot. It's better than any alternatives I've seen which require too much geek knowledge to operate. (Admittedly I've never seen how MacOS X handles updates.)
-Teckla
hmmm... (Score:4, Insightful)
I have had windows update tell me that i'm clean, when i've only just done a fresh install, but i don't take it personally, you'd only complain if it examined every bit of your disk to ensure that it got it right... make your minds up people!!
Re:Then work on an alternative... (Score:4, Insightful)
Re:I don't trust Microsoft... (Score:3, Insightful)
BS BS BS BS BS BS BS
Yes, their patches do on occasions break things. Not defending that, they need to be more careful sometimes...
But "MORE OFTEN THAN NOT" is FAR from the truth, and I am sure you know this. But, with your M$ $ucks patch sewn directly on your forehead, you kinda hafta make remarks like this, right?
On the few occasions things break they are rarely of the "blow up the server" variety, and MORE OFTEN THAN NOT *grin* they are of the "when the stars align" kind that you HEAR about in bug reports but don't experience first hand.
Single Point of Failure (SPOF) == Bad Thing(tm) (Score:5, Insightful)
From the article:
we no longer even need to make that visit manually, we can trust that Microsoft will supply us with a properly tested security patch within 24 hours and patch our systems for us
What follows is not MS-bashing.
System security (and other functions) should not be left to a laissez-faire, set-it-and-forget-it sort of mechanism. The administrator is responsible for applying the patches, manually if need be, and should be diligent enough to determine whether all requisite patches are installed even when using an automated method like Windows Update. Yes, that includes apt-get, RHN, up2date, and others.
I believe it also behooves the administrator to conduct independent testing on-site: there have some notable examples of patches getting out the door that caused as many problems as they solved. (Yes, I'm thinking of SP4 for NT 4.0. Still not MS bashing, though.)
Trust, but verify.
Don't trust it? Don't use it. (Score:5, Insightful)
Re:I like Windows Update (Score:5, Insightful)
Re:In case of slashdotting, (Score:0, Insightful)
Maybe not... (Score:5, Insightful)
Let me put it this way. Since the inception of Windows Update millions of computers have been infected with Trojan's that are today allowing individuals to conduct en-masse DDoS attacks. Read that how you want, but its a fact. Here's another. Since the inception of Windows Update Microsoft has gone to producing patches almost every week. Few if any business' have found Microsoft trustworthy enough to permit automatic updates
Many people will also tell you that a false positive is far worse than a false negative. For example, if Windows Update is misconfigured and tells you that you're up to date when you're really not, that's arguably worse than not being up to date and knowing that you're not up to date. (Because in the latter situation at least you can do something about it)
Even if technically windows update is better than nothing, it's utterly pathetic that this is the best one of the richest and most powerful corporations on the planet can do for their customers.
Hm (Score:2, Insightful)
I mean, if my headgasket in my GM blows, I don't go to Goodwrench with the schematics for a new design.
Re:it's better than nothing (Score:3, Insightful)
A different way? (Score:2, Insightful)
I think its a win/lose/lose type of situation.
Re:Trust? (Score:3, Insightful)
That's pretty funny.
Wrong (Score:2, Insightful)
I can't think of a reason why any business or corp at this point is patching their Windows boxen directly from Windows Update. SUS make it a whole lot more controlled, reliable and faster.
But it is a consumer OS (Score:5, Insightful)
Re:Don't trust it? Don't use it. (Score:3, Insightful)
Windows Update is supposed to be the "magic button" that makes your computer safe. But it doesn't work, and there are still lots of alternatives that will keep your computer safe. You just can't expect it to be as easy as pushing a button. That's why they pay Systems Administrators.
The thing I don't like about Windows Update (Score:5, Insightful)
I just experienced this two days ago. My friend had me reinstall XP on his laptop so I started with a disc that had XP SP1 included. Now considering the huge list of known problems SP1 causes both he and myself were happy with how the system preformed after install. It seemed snappy and worked well. But then after I ran windows update and pulled down like 15 security updates, boom instant slowdown. I'd say its about 15-20% slower now. I might as well have pulled out his PIII900 and dropped in a PIII600. (And yes I specifically avoided 811493)
When will MS stop having to reissue patches and stop slowing down and screwing up systems because they can't figure out how to make software with some decent security built in? I mean screw the security track record of other OS's, Microsoft is the one with 40 billion in the bank. They are also the ones who still don't get it and are just now telling their programmers that security needs to be considered when designing software. For about the fact that OSS exists, I still can't believe people can people can have faith in a company like that.
Re:Don't trust it? Don't use it. (Score:4, Insightful)
It's a completely different situation than not using NVidia.
Re:I don't trust Microsoft... (Score:2, Insightful)
Lucky you.
You can keep your faith in M$ but, speaking from experience, when you install a seemingly innocuous update then get BSOD on boot and have to rebuild the box from scratch you tend to loose your trust in them. It is probly half the reason that people don't patch and end up being hacked. So either way you loose.
That's it -- just dismiss all faults as MS bashing (Score:1, Insightful)
Re:I like Windows Update (Score:4, Insightful)
>I'm sorry, but the separation of download and install steps is a good idea.
Two users who disagree. Solution would be to make the behaviour configurable then, yes?
Cheers,
Ian
Re:Trust went out for me.. (Score:2, Insightful)
Re:Maybe not... (Score:5, Insightful)
So if that's a problem with Windows Update, perhaps that is why many companies still don't trust Open Source. The only difference here is that we don't see the source code. I don't read the source anyway, so I'm not losing anything :P
Re:I like Windows Update (Score:1, Insightful)
Re:Summary (Score:2, Insightful)
Trust is not about users trusting the corporation.
The trust they want to implement is the other way around.
Trusted computing involves the hardware running only "trusted" code, the screen showing only "trusted" documents. And knowing is half the battle.
Re:MOD Parent Up = +7 Anti-MS (Score:2, Insightful)
Thats funny, and probably true (and definately shouldn't have been modded to -1), but thats missing the point a bit.
Its not a bad thing to be about profit, but it is a bad thing to put profit ahead of other concerns, especially when you are an industry leader. I think that the outcry would be the same way if Ford knew that a part was faulty, and they supressed the knowlege or downplayed it in the press. How about Boeing? Should either of these companies put their corporate reputation and profit ahead of safety? Of course not.
Now, you might say, whatever - nobody ever died because of a Microsoft trojan horse. And I would agree - but they have caused hundreds of millions of dollars of damage and hundreds of thousands of wasted man hours - all beacuse they are unwilling to reveal themselves for what they are - human.
First, they need to admit that they make the occasional mistake. Secondly, they need to make an easy and trustworthy way of recovering from those mistakes. And thirdly, they need to make it seem like they care more about about the security of their existing customers than trying to gain new ones. Its that easy.
Re:I don't trust Microsoft... (Score:1, Insightful)
problems and solutions (Score:5, Insightful)
Re:In case of slashdotting, (Score:5, Insightful)
In Open Source, complaining like this might be frowned upon sometimes. After all, we understand that not every OSS developer works for IBM, and has time and resources to fix every bug.
However, this is commercial software, and closed source to boot. Why should anybody solve Microsoft's problems? Isn't that why people pay for work being done for them in the first place? I think he's doing pretty much the best thing he can do, complaining in public. That's the one thing that seems to work pretty well to get the attention of large companies.
Re:In case of slashdotting, (Score:3, Insightful)
Re:Maybe not... (Score:3, Insightful)
I agree with you on the false positive scenario except that you've left out the most likely case without Windows Update, a nothing, because without Windows Update right there quick and easy to use, most people just wouldn't check at all.
So I stand by my first statement. It's better than nothing.
Re:turn it off (Score:2, Insightful)
Re:Maybe not... (Score:3, Insightful)
Are you drunk?
Picture this: You are the sysadmin at a company that runs its business all online, doing thousands of dollars of business per hour. You have a farm of 2000 servers running a custom back end for all your web services. The weekly patch comes down from Microsoft, it's time to update it. Again. Just like last week and the week before. You go down to the colocation facility late that evening, and apply the patches because that's the corporate policy. A few minutes after applying the patches you notice that your company's custom apps are acting.... odd. You call one of the programmers who works on the app, and he sees the same thing. Soon the whole team is conferenced in, and the consensus is that the patch screwed up the company's app. Rollback! goes the cry and hue!
You attempt to rollback. It doesn't work.
Fuck.
Your options at this moment in time are:
a) Try to patch your internal app to work around what Microsoft's wonderous patch has broken
b) Spend the next 3 hours (per machine) reinstalling and restoring from the tape backups to be ready for the opening rush of business by the next morning. No, you are not getting paid extra for this.
c) Clench your collective buttocks and see how badly this fault affects normal business, all the while praying the Overlords at Microsoft release a patch for this patch real soon.
d) Pick up one of your servers, hurl it through the window in the colocation facility (on the 21st floor) and jump after it to the blissful eternal night where there is no Microsoft (that we know of).
Answers to a couple of questions that might come up:
Q: Why aren't you testing all of these in a test lab before going live?
A: I don't know. Probably because we spent all our damn money we would normally use to hire a regression testing QA team on server licenses instead. Call me naive, but when we're paying multiple thousands of dollars per server on software licensing, is it too much to ask that the shit doesn't require us to hire a QA team to constantly regression test the effects of Microsoft's bug fixes? Since this is something we're trying to find the budget for, apparently it is too much to ask.
Q: Why are you using Windows 2000?
A: That's what our app is developed on. And continues to be. You don't throw out 15 man-years of coding on a whim to switch to Linux just because this year it's "finally" being seen as ready for the enterprise by enough people in our company. We have Linux boxes a go-go here. Just not doing this.
SP1 hogging memory (Score:3, Insightful)
But to you, it suddenly becomes "XP SP1 hogged memory."
Next.
Re:Trust? (Score:4, Insightful)
Re:Trust? (Score:3, Insightful)
What an odd thing to say. You don't trust them in general buy you trust them in this particular case? Why? That's like saying "I don't trust that convicted child molester living across from me but I'll let him babysit my kids because nobody has proven he will abuse MY kids".
Re:Maybe not... (Score:3, Insightful)
E.g., Red Hat may drop support of 6.2, but they don't force you to upgrade to 9.0... they don't even force you to upgrade. Now eventually some hole may be found, but in the intermediate time period you can have been studying OS versions that have been out for a year, looking for what seems most stable and appropriate for *your* needs. (For that matter, you could be using Debian stable. Or an appropriate BSD... I think Debian is the better choice, but your needs may well be different from mine.)
Re:Then work on an alternative... (Score:3, Insightful)
I haven't checked lately, but it's very possible that there is something that prohibits it.
Microsoft's patch files are, after all, their own copyrighted property. Redistribution would be illegal unless they've given you specific permission. (Many software companies explicitly deny this permission, even for products which are free to download. Sun's JDK for example)
There are other legal pitfalls- reverse engineering, for example, might be required to check if a patch is needed. (You'd be writing code to check if there are security problems, which edges towards violating the DMCA or at least a EULA)
And anyhow, while some Linux developers are happy to do free work for IBM, you're less likely to find open source coders willing to put in time to fix Microsoft's oversights- especially for a field as unglamorous and time-consuming as patch distribution.
Re:Maybe not... (Score:3, Insightful)
Perhaps the SQL worm rings a bell? The discussion about that on Slashdot was full of guys like you saying "Geez, anyone who didn't patch right away and got hit by the worm deserves to be fired!"
Patch when it's going well in case of the next wild exploit, or don't. Can't have it both ways.
Personally, I'll go with the lesser evil and patch. Better that than be part of the mess when the shit hits the fan. Unless of course you like the idea of your boxes being part of a half million zombie machines that attack the DoD or something...