Slashdot Mirror
Microsoft Says Two Basic Security Steps Might Have Stopped Conficker
coondoggie writes "If businesses and consumers stuck to security basics, they could have avoided all cases of Conficker worm infection detected on 1.7 million systems by Microsoft researchers in the last half of 2011. According to the latest Microsoft Security Intelligence report, all cases of Conficker infection stemmed from just two attack methods: weak or stolen passwords and exploiting software vulnerabilities for which updates existed."
Troll much? Windows has nothing to do with it when you set all of your passwords to "123456".
So basically they're saying if you had better passwords and applied patches, you'd avoid security problems?
Nice to see MS on the cutting edge of security research.
I want to delete my account but Slashdot doesn't allow it.
We have better authentication methods, we are just not bothering to deploy them. How many times do passwords have to fail before we acknowledge that they do not provide the sort of security that we need?
Palm trees and 8
It's not my fault!
True, but there are targeted attacks even in the Unix world, and if you don't keep it up-to-date, you could be owned by one of them
Fanboy? No, I actually run Mac and Linux at home and I program cross platform at work. The fact that Conflicker happened to be for Windows has nothing to do with this. Running old software with weak passwords is a recipe for disaster on any existing OS.
Which wasn't even properly disabled when you tried to disable it through the UI in Windows. Who were the idiots not following security best practices when they came up with that idea? Infected flash drives and non-disabled autorun were the main vectors for Conficker around here.
Please name a Unix based attack that is equivalent to the malware being discussed.
CAN-CON 2019 - Ottawa's only book oriented Science Fiction Convention! October 18-20, Sheraton Hotel, Ottawa, Canada h
The assumption here is that an attacker choosing the easiest way has no other route. It would be safer to say that the route used by the worm would have been unavailable if basic preventative steps had been taken.
It's like the old joke. "Ever wonder why whatever you're looking for is always in the last place you look?" "Well, sure, once you've found it, why keep looking?"
Microsoft seems to think the authors would have stopped looking without finding an exploit route. Instead, they found one, and stopped looking.
For example E-trade will give you the RSA key fob. Am I supposed to get a dozen key fobs from each of my bank, brokerage, mutual fund, anf 401-K administrator? Schwab would not let me use special characters in passwords. I think they also have a ridiculous 8 char limit. In this day and age where GPUs are being used for dictionary attacks? 8 char? Fidelity wanted an all numeric password because they wanted the phone based log-in used by their older customers to work in web too. On top of all that they have the password reset procedure which asks for stuff that you can find on the facebook profile.
Then there are idiotic Paychex which will lock you out after two failed login attempts. There is this site securetransfer.com that requires some 16 char password with at least two capitals two numerals and two special characters to get 100% strong password quality rating. Then there are clueless admins who tell you "never write down the password". Hello! Is there any end to this password madness?
Why can't they give me two levels of access? Read only access that lets me see account balances and verify that the check has cleared. And the write access that requires one more password that allows me to transfer funds and trade securities. May be even a third level password to send cash out of that institution to outside.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
Yes, because it's completely impossible to turn that feature off. Oh wait...
http://windows.microsoft.com/en-US/windows7/Turn-automatic-updating-on-or-off
If you don't want them "forced down your throat", maybe you should change the setting to instead notify you that they exist and then let you pick and choose which ones you want to install as well as those you want to ignore permanently? How is that any different from any of the automatic update services in Linux distributions bugging you to update and you continually ignoring them?
The difference is that unless it's a kernel update Linux doesn't really need a reboot on update.
I am Bennett Haselton! I am Bennett Haselton!
It's nothing like the Windows situation where you get a bag of critical patches forced down your throat every Patch Tuesday, and then your Windows box loves to reboot right in the middle of whatever you are doing. Sheesh.
1) Just as a point of clarification, Patch Tuesday is only once a month. And there's usually only about a dozen or so, only some of which are genuinely "critical". Obviously that varies though. 2) Windows Update has been a lot better for years, ever since Vista. There's nothing wrong with it now. You might be able to complain about the default settings, but they're right there and they're pretty straightforward. If you're logged in and it's set to restart automatically, it prompts you to restart or postpone it. And, obviously, you can shut down the automatic reboots or the automatically downloading/installation of updates. Besides, since moving Windows Update to an actual program after XP, there's also been a lot fewer updates that seem to require restarts. With XP, it seemed like you had to restart every single time you ran updates. Vista/7's a lot better with that.
Microsoft gets to say, "hey we patched that before it was a problem". That's an unusual position for them to be in.
It's actually not an unusual position for them to be in at all. The vast majority of major Windows worms exploited vulnerabilities that had long been patched.
I don't always use unix-like operating systems; but when I do, I prefer FreeBSD.
It's nice to keep telling people "you wouldn't have the security issue if you did all the updates right away". But to that, I'd like to tell the OS developers something else:
You wouldn't have the concerns about unpatched systems if you designed the OS so it could apply the downloaded updates without requiring system reboots!
And yes, though I'm not a software developer, I do know a little bit about this, and why it's a "tall order" (core services you can't just delete and replace with updated versions while they're in use, etc.). But I guess I'm saying this doesn't seem impossible to overcome, if someone wanted to make the functionality a priority in a new OS's design?
Unless we reach that point, people will always be delaying installation of new updates because it interferes with work they need to get done, or they're afraid an update could potentially break something they rely on and don't have time to deal with, if it goes wrong. System patches/updates need to become a less intrusive, more seamless process -- and one that can easily "roll back" any new update that turns out to cause issues. It should automatically notify the developer when this happens, and should flag the problem update so it doesn't get re-installed (but subsequent, supposedly corrected versions DO get installed ASAP).
With today's multi-core CPUs, maybe it's even possible to design systems so two instances of the OS/application environment can be run in tandem during an update process? Hand off the running processes to a parallel copy of the current environment, invisibly to the user, when an update is about to take place. Then patch the first environment, which now has no "core services" in use by apps anymore, and shuttle the apps back over to the patched environment when it's ready?
For this to work, companies would first of all have to agree to run their update process through said package manager. You don't think this will ever happen, do you?
What bugs me about Windows is that there is very often no way to do an unattended update at a certain time for many "packages". Windows being the notable exception. The average Windows day for the average customer runs a bit like this:
"Ok, I'd like to play a game. Let's double cli... huh? Oh, Acrobat update. Ok.... yes, accept license... wait ... download patch, watch download bar move... installing... watching bar move ... ok, we're set. Now lemme... huh? Oh, virus killer. Ok, 'tis important, go ahead and update yourself. Yes, license agreement... waiting for download (because experience taught us that you better NOT try to do anything as system critical as starting a game while something is being patched. Could upset the copy protection trojan). Huh? Failed? Oh, because the Acrobat update didn't finish yet. Ok, it's finished now insta... restart."
"And we're back after the break. Now, for the antivirus. download ... update... huh? New version? Ok, install it. Yes, I agree with the license... installing... reboot."
"Finally! Ok, first of all, let's take a look at some porn. Open Browser... oh, new version? *sigh* Ok, download and install it. ...waiting... Ok, now... huh? What happened to my plug... oh. Of course. Incompatible. Fine, but I'm not going to visit any porn pages without a decent ad blocker, so first of all, update the plugins."
(half an hour of browsing, finding them, or not finding them and searching for a replacement later ... And another few minutes later including washing your hands...)
So. Game time! Fire up Steam... updating... Ok, restart steam... While it's doing that, let's start Teamspeak... Oh. Updating... must be patch day all over the world...
Finally a good game of $whateverfps. Huh? Patch? I don't wanna, not again! Oh, no multiplayer without, huh? Ah, anti cheat stuff. Ok, make it so...
And so on, and so forth. THIS is what actually bugs me about Windows. The piecemeal updating process. You can't just keep your machine running to have it update its stuff and actually, you know, USE it when you are sitting in front of it. It seems to be critical to steal the user's time and show him that they actually patch their half baked software.
And it's not like the software (and its patchers, launchers and oh-so-important taskbar tools) wouldn't run anyways and could technically do a daily check for updates. Dear Adobe, care to inform me why you insist that your launcher is running (and turning it off only means it gets reinserted into the Run key as soon as I dare to open an Acrobat document) and steals my ram for zero return, yet STILL require me to be present for every damn update you might want to run? Why is there no option in Steam to automatically patch and restart Steam if I'm not currently playing a game?
Rolling that all into a single package handling goodie would be a blessing. And MS actually manages to do just that with their updates, the kicker is that of all the various companies that have their fingers in my system, MS bugs me the least!
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
Severed and sheared? Your workplace sounds way too violent.
It doesn't mean much now, it's built for the future.
It's really hard for me to say that, but getting rid of Windows isn't going to do jack. Idiots using computers will be vulnerable to malware, no matter what kind of OS they use. Unless the OS is secured away from its user, there is no safety if the user himself is the biggest security hole.
The key to the whole issue is the Dancing pigs problem. In a nutshell:
"Given a choice between dancing pigs and security, users will pick dancing pigs every time."
People don't even notice the warning message, and they don't care. Why? Because they got way too used to it. UAC pops up and wants you to say yes to something, and people will click yes without thinking what's going on. Why? Because they learned the wrong lesson. They lesson they SHOULD have learned is that this window tells them to go and think whether what they are about to do should really require administrative privileges. Should displaying some childish webpage require the rights to dig into your system's bowels?
What they learned is "if I click no, it does not work". That's pretty much it, this is the way people work and think. They don't WANT to know what this window means. For them, it could as well not exist and if anyone ever tells them how to turn it off (and yes, you can), they will without thinking twice and be grateful that they got rid of that nuisance. And, bluntly, it doesn't make a lick of a difference for them anyway!
Why the heck would this be different with, say, SE-Linux? You know SE-Linux? Allegedly one of the more secure and hardened Linux flavors in the world. Hand it to Mr. Moron now using Windows 7 and it will be "pwned" in minutes. Allow me to illustrate.
Let's assume he is using Linux, even properly configured by a good friend of his who made the horrible mistake of telling him the root password. In comes my trojan, disguised as some kind of, say, torrent speed enhancer. I'll even be blunt and forward in the reasoning just why he has to install it as root.
"The software needs elevated privileges to install and properly configure the device driver needed to establish a secure connection with the controlling server to maximize the success and streamline the process. This also allows the software to work without any user interaction necessary, you will not have to enter the password ever again for this software to function properly"
In short, let me install my rootkit and hook up a connection to my bot herder server.
What will Mr. Moron read in this sentence. He doesn't understand it, at least not all of it, but he knows a few words out of that and here's what he puzzles together from this:
"The software ... technobabble ... install and properly configure (ok, it does that by itself, I guess, but only if I type in the password. If I don't, it probably won't work properly)... more technobabble ... server (server is good, I want to connect to one. I think) to maximize the success, streamline process (yeah, I want that!). No user interaction necessary later on. Never have to type the password again (great, so just once and then it works on its own. 'k, no problem, once doesn't count, right?)
He WILL hand over his credentials. Without thinking twice. And he will have forgotten about it before the trojan makes his first report to his controlling server.
It doesn't matter what system you give him. Security is the minimum of the system's capabilities and its user's capabilities. Not the average. The minimum thereof.
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
Thats a false argument. You give me equal amounts of clueless users using Linux as they are with Windows and I'll name one.
The vast vast vast majority (I'd say 90+%) of Linux PCs are (1) servers that are administered professionally or (2) locked down cell phone OS or (3) desktops that geeks use. There is no way you're going to be in the same situation as Windows is with that kind of demographics.
ELF, ld.so, and dynamic library versioning pretty much eliminated that. Or are you one of the few that actually manually removes an old C library version and then rebuilds every single executable that complains it can't find the old version?
Again. Just in case I didn't make my point clear.
The user hands over the password.
It's not a trojan reading the file where the password is stored. It's not a hacker getting in from the outside using some supersecret backdoor account. It's not any kind of hack whatsoever. How the heck do you want to keep a password secure from its rightful owner and user?
The USER is the problem. Not the system. And unless Linux has some magical ability that I didn't notice yet, namely the ability to know what the user WANTS, instead of just what he DOES, there is exactly zero chance to protect the password. No matter the system.
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
MS is in a bind here. They are very much aware of this problem, but there is very little they can actively do against it.
It's not even MS that is the problem here, it's the way some companies (notably game companies) abuse the system and don't write to spec. In Linux, you get ravaged (to avoid a less pleasant word) if your software required more privileges than it absolutely minimally needs, and you better have a GOOD reason to ask to run as root. Hell, most packages say explicitly that you should NOT run this as root.
It's exactly the other way 'round for MS Windows. With both, old legacy reason and newer, at least as bad reasons.
The legacy reasons come from the times of the Win9x systems who arguably had zero real protection. Likewise, it didn't matter just what Registry tree you cluttered with your keys. And because it's easier and works for all users to simply slap it into the HKLM tree instead of the HKCU (aside of other, more serious, problems that you have to take into account when using HKCU), software creators didn't even think twice before sprinkling the Registry liberally with their crap. Of course, this flies right in the face of anything resembling security where HKLM or even HKCR are off limits for "user" privileged accounts. So every time this legacy junk was supposed to run, UAC throws a hissy fit.
The less acceptable reason and the one that irks me way more is that the various DRM schemes and anti-cheat crap make games require administrative privileges, not only for installation (where I could at least accept that, due to installing a device driver, these privileges are required) but also to run them. Again: To run a stupid, insignificant game, you have to bring out the big admin guns. And this is simply NOT ok.
But there is very little MS can actively do against that. As long as people buy those games despite the need for admin privs, companies will continue using DRM schemes that don't give half a crap about the system's security. And as long as this is the case, MS cannot do anything about it. What should they do?
As soon as a program requests permissions that can somehow harm the system, a sensible security watchdog function should report that something is happening that could be damaging. Else, what is it good for? The security of the system is the security of the weakest link. One link broken, the security breaks down. You can't simply "not ask just this one time". If you do that, disable it altogether. But if it really asks every time something could possibly be amiss, you get what UAC is today, along with its "allow and deny" jokes.
So please tell us, what should MS do?
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
Sure, sure, blame the users again, Microsoft.
How about educating them for once? You own, according to some metrics, 90 percent of the desktop market. Your operating systems in retail boxes don't even come with quickstart guides to basic security. No, you just leave your users to flounder about without any guidance at all, and if they want it, they have to pay extra for it.
At least when I was paying for boxed sets of SuSE Linux, it came with two well-written manuals, a user's manual, and an administrator's manual. I suspect that boxed sets still include these. It was in the grand old tradition of "when you get this software, we'll give you the manual too" like what you got when you bought DOS or CP/M.
But these days, I guess that user education is viewed as "intimidating" to users, because *shock* *horror* computers might be revealed as the complicated, useful, and powerful devices they actually are and heaven forfend users get any ideas beyond clicking on the pretty pictures. Microsoft does its damnedest to not give the user *anything* that might resemble common sense lessons in security.
There is a lot of energy pointed at the education of developers, but none that I can see at day-to-day users from Microsoft.
I just dealt with a user who has become so paranoid, she considers technet.microsof.com "foreign" because she's been so abused by the utter lack of guidance in the past with computers that she can no longer tell what's legitimate or not, wrt software. I was merely pointing out a sysinternals tool. This makes me a sad panda, and I don't blame her. I can't. Because I've seen it too many times to think it's just "dumb users" anymore.
Microsoft's blaming of the user is utter bollocks. It is entirely their fault now.
Yes, this makes me mad. Deal with it.
--
BMO
Where do you think the term "Root"kit came from?
Before NT Unix was the laughing stock off security seriously. Like Windows it is also written in C and uses the same apis for buffer overflows, stack over runs, and other crack attacks.
My old World Almanac from 1990 had an editorial on the first ever Worm which nearly took down the internet. Hint ... it was all Unix based.
You hit the nail there.
ASLR and the other OS protections are untouched because most corporations still use XP and a 10 year old kernel. The reason most software doesn't use these things and tap into them is because they wont run on XP. Corporations wont leave XP because software doesn't use things and tap into them. Cost savings are on top of this.
This is a great reason to upgrade to Windows 7 and keep your systems patched. This was totaly preventable and IT departments got what they deserved for their short sightedness on only cost savings.
http://saveie6.com/