Stuxnet May Represent New Trend In Malware

Trailrunner7 writes "As more information continues to come out about the Stuxnet worm and the vulnerabilities that it exploits, it's becoming increasingly clear that this kind of attack may be a preview of the attacks that are likely to become commonplace in the months and years ahead. The most interesting aspect of all of this is the fact that the attackers behind Stuxnet clearly knew about the vulnerability in the Siemens WinCC system before the malware was written. That implies the malware authors had some advance intelligence about the configuration of the Siemens software and knew exactly where there was a weakness."

Comments

Uh - what?

[Anonymous Coward]

The article that the summary links to in support of the idea that the attackers had inside information is actually about a hardcoded password that existed for *two years* before the vulnerability was found. The article argues pretty strongly that security through obscurity is no security at all and makes no mention of anyone having inside information - how can you get it so wrong?

Re:Uh - what?

[v1]

I see the article boiling down to a different point -- should vendors be held liable for exploitation of a bug that was brought to them some time ago? Article says they knew about a hardcoded pw two years ago and sat on their thumbs, and then it questions whether this is negligence. There is no question. That is negligence, they will be sued, and they will lose.

Since we keep seeing things like this come up over and over, it seems reasonable to assume that companies like this simple consider things a "calculated risk", and determine the chance of being caught x the cost of being caught is less than the cost of fixing it, and so they do nothing.

The only way to fix this is to increase the average cost so that it becomes greater than the cost of fixing it. To accomplish this, customers should be able to sue vendors that have been informed of critical security flaws in their software that have not fixed it in a timely manner, and there should be specific laws on the books for fines to be levied on companies that manage to not get sued until their refusal to fix their bug is being exploited and harming their customers, to make the resulting legal actions much more expensive than simple lawsuits from individuals. (why aren't these things considered "class action"?)

Re:Uh - what?

[LeDopore]

Article says they knew about a hardcoded pw two years ago and sat on their thumbs, and then it questions whether this is negligence. There is no question. That is negligence

Not always. Some control systems are run on a dedicated computer without Internet access. Some control systems need to have little downtime to avoid serious consequences. (Some manufacturing plants or refineries have razor-thin margins - an extra 1% downtime could mean the difference between profit and bankruptcy.) In cases like these, if a hard-coded password means a faster system recovery, it's the right choice.
If I had software on my desktop system with a hard-coded password, I'd be justifiably pissed. However, for some industrial applications (including some SCADA installations) , the simplicity of not needing to enter a unique password plus a physical air gap of security trumps a forced-unique password with only digital security - particularly if that digital security is Windows-based (where adding a keylogger would have resulted in almost as bad a p0wnage as what Stuxnet already has)!

Re:Uh - what?

[Ephemeriis]

Not always. Some control systems are run on a dedicated computer without Internet access. Some control systems need to have little downtime to avoid serious consequences. (Some manufacturing plants or refineries have razor-thin margins - an extra 1% downtime could mean the difference between profit and bankruptcy.) In cases like these, if a hard-coded password means a faster system recovery, it's the right choice.

So, why not have a password that is generated in some known way?

The HIS system where I work has a "daily password" - it changes every day. That password is necessary to conduct some operations. Folks who need to conduct those operations know how to look up the daily password. They do so, then they have that password to hand out to whoever needs to do stuff that day. And the daily password becomes useless the next day, so you don't have to worry about it being abused.

The POS system I used to work with had some kind of dynamically generated password. If you had to call technical support for something they'd have you read off some numbers on the screen, and they'd give you back a password to get into the register's internals. Again, it isn't static so it can't be abused for long. But it is generated in a known way so it can readily be obtained.

Seems to me that this would have been a better way to do things.

However, for some industrial applications (including some SCADA installations) , the simplicity of not needing to enter a unique password plus a physical air gap of security trumps a forced-unique password with only digital security

"Air gap" doesn't mean much if you're just using some kind of removable media to transfer information from the insecure world to the secure world, instead of CAT5. If you aren't somehow protecting access to that removable media, your air gap gives you no additional security.

It should be genuinely impossible for anything to auto-run on removable media. Only allow media in your own, special format. Or only allow specific file types to be accessed or imported. And put some kind of password on the media access portion, to make sure only folks who know what they're doing are accessing it.

If you're letting anyone transfer anything on a USB stick, you may as well plug the machine into the network because your air gap isn't doing you any good.

Re:Uh - what?

[KDR_11k]

It's a result of strong competition, except for price collusion there isn't really a way out of that situation once you and all competitors have driven the margins that low.

Re:Uh - what?

[Runaway1956]

"(Some manufacturing plants or refineries have razor-thin margins - an extra 1% downtime could mean the difference between profit and bankruptcy.)"

I think you exaggerate. If not, then the executives and other decision makers are incompetent boobs, and the company SHOULD go bankrupt for hiring them.

I'm a maintenance man at a manufacturing plant. Our management is only marginally competent - but we have enough redundancy to see us through when something goes down. Sure, it HURTS when a computer dies, and the vendor won't honor the guarantee unless we send the computer back to the factory. But, bankruptcy? Come on . . . .

Re:Uh - what?

[HAKdragon]

There aren't any commas, either.

More common?

[Spad]

Given that we have absolutely *no* idea how many similar attacks have been conducted in the past against really "niche" applications like this without being detected, I think it's a little naíve to assume that this is the start of a new trend.

We find out about most malware because it's so widely targeted and so many people are affected by it, but when you're targeting your malware at a handful of companies and probably directly delivering it via email or physically ("dropped" USB stick in the parking lot) with the aim of keeping it undetectable for as long as possible, it makes it much more difficult for the targets and security researchers to even know it exists.

How can he get it so wrong?

[NotSoHeavyD3]

Umm, you do realize this was something posted on Slashdot, right?

Who could have guessed?

[briuq0ah]

Seriously? The attackers knew about the vulnerability before they wrote something to exploit it? I never would have guessed

SCADA frustrations

[brxndxn]

My career is in industrial automation - and I am an IT guy who 'gets' both sides of things. There are not a lot of people like me and I constantly face an uphill battle when I try to explain computer security to people or try to explain why certain things are much more complicated than they believe. For example, you have an industrial network that is completely unnattached from the corporate network that is used for automating an exothermic chemical process on a large scale where you cannot just 'hit e-stops' and safely shut down the process. If you lose 'visibility' on the process at any time, there is potential for an explosion or chemical release. They think they're immune to viruses and they do not run virus-scanning software (imo, usually a good thing in an industrial network) so they do not even bother to completely lock down the computers. We're talking Windows boxes where everyone knows the admin password. After a virus or two, they usually pay me to lock everything down and put the operators on limited profiles. Then, the white-collar management wants to be able to connect into everything to see what is going on. Suffice it to say.. it's a damn headache. IT doesn't get it and the plant managers don't get it.. And usually one wins out over the others. If IT wins, expect a plant to randomly shut down because they push an incompatible Windows patch. If the plant wins, expect a laughably insecure network where an operator charging his cell phone can take the whole network offline.

Basically, if you ask an IT guy 'What is security?' it will be a lot different than an industrial plant manager's response. An industrial plant manager will say a SCADA system is most secure if the people on site always have control over the plant. If a man has his hand caught in a machine, should another person at the plant have to login to a terminal to turn the machine off?

I'm frustrated by this virus, though, because from what I've seen, there has been NO utilities released to detect if you have it. I have seen abnormal activity on multiple HMI computers and the people in charge of maintaining them plug their thumb drives in randomly thinking as long as their laptop doesn't detect a virus on it, they're safe. At least conficker was obvious to detect on a thumb drive or running computer.

If there is a utility, can someone link to it for me?

Re:SCADA frustrations

[Anonymous Coward]

Full Ack.

I have worked for a well-known company who builds large plants for various industries (including food processing). The SCADA systems they set up were a real nightmare. Most plant-controlling computers were directly connected to the internet (no NAT), not even a personal firewall was used. Some had even activated the Windows default shares (C$ and such). The computers were never patched, and the software they used for remote administration transmitted login data unencrypted.

The people who configured and developed/extended the SCADA systems were mostly engineers coming from the machine industry or electrical engineers. They didn't know much about IT, and even less about IT security. I was never in a position to look at the source code (I was working for the internal IT department), but I bet it was full of holes.

Sometimes we had to fix software problems for customers, on plants that were installed a long time before I was working at the company. Some of these plants had computers still running on Windows NT 4.0 - because the SCADA system they used didn't work on any newer OS and they didn't want upgrade it (cost). I was amazed that this thing was still running (more or less).

I bet that even a medicore black hat could take most of these plants down within hours - this could be a real threat.
I guess the only reason this hasn't happened very often yet is that it doesn't give the attacker much profit.

-- AC

Re:SCADA frustrations

[leuk_he]

NO NO NO. installing a cocktail of AV software is NOT the answer in a system that has to do 27/7 operations and has to be kind of real-time responsive. What you do not wat it that it gives a false positive on an essential program that controls the plant. you do not want to inititiale a scan at the moment 2 reactive componentes are mixed.

as for the tools, look for
  Right click My Computers > Properties > Hardware > Device Manager:

            - Go to View > Show Hidden Devices
            - Go to Non-Plug and Play Drivers

        Disable both MRXNET and MRXCLS:

If that is the case you might want to look for some portable virus scanner...

Re:SCADA frustrations

[Metroid72]

Sounds like your organization doesn't have the right support for security (most likely they don't understand).
You need to engage in a business impact analysis exercise and see what comes out of that.

That will be input for your policies, procedures, standards, etc.

Re:SCADA frustrations

[tchuladdiass]

I've got a good way to deal with many root kits. What is the one thing that a root kit does well? When you read an infected file, it will give you the "clean" file's contents (by intercepting the OS read system call). So the way to deal with them is simple: You're enemies strength is its weakness.

While the OS is running (with root kit), make a copy of all OS files (c:\WINDOWS, system32, drivers, ...). The root kit will make sure you have clean versions, since that is what it wants you to see. Afterwards, boot off a live Linux CD, then cmp the OS windows directory with the copy you made. Any executable file that is different is likely to be a hiding place for the virus.

Now it is possible for the root kit to still write out infected files when you make copies of them, but this process has worked on the last several infections I've had to clean up. Follow up with a regular virus scan afterwards to catch any registry shenanigans and you should be golden.

Re:SCADA frustrations

[Runaway1956]

You seem to have read AC's post differently than I did. He didn't suggest that you INSTALL an AV cocktail on your machines. He stated that a cocktail of security applications are necessary to deal with an infected machine. That Geek Squad disk is exactly that - a bootable disk with quite a variety of utilities, suitable for dealing with a wide variety of problems.

I can't remember offhand, but I think there are six different AV's on the disk, each of which automagically updates itself when the utility is called. And, each of those AV's is on the disk, because it has proven superior to any of the other in certain situations.

Your advice is sound, you just seem to have misinterpreted what AC said.

Re:SCADA frustrations

[Runaway1956]

"Some of these plants had computers still running on Windows NT 4.0 - because the SCADA system they used didn't work on any newer OS"

Guilty as charged. Yep. A hard drive died recently, and a machine that is worth half a million sits idle because of it. "Don't we have a disk image? I can get this thing running in a few hours, if I can run to the store for a hard drive!" "Disk image? What the hell is THAT?"

Phhht. No backup, in any form. And, this expensive machine sits idle due to the failure of a ~$50 component.

Fortunately, MOST of our equipment runs on Linux, and MOST of our equipment just runs and runs and runs.

Re:SCADA frustrations

[Anonymous Coward]

what a nice buzzword

business: $$$ for me
impact: a forceful, exciting word
analysis: smartness, i'm one of the smartest people on the planet
exercise: what i should do more of to keep my tight body

Re:SCADA frustrations

[Jaxoreth]

NO NO NO. installing a cocktail of AV software is NOT the answer in a system that has to do 27/7 operations and has to be kind of real-time responsive.

That's why most shops don't offer more than 24/7 uptime. The three extra hours a day is plenty of time to run AV software.

Fail

[PPH]

Right click My Computers > Properties > Hardware > Device Manager:

That's Windows you're working with. And if you are using Windows in an industrial environment, you've already screwed up. Forget about the usual /. nonsens of Windows Bad, Linux Good. Actually, you'll need a special RT version of Linux. According to Microsoft licensing terms Windows is NOT suitable for use where life or property loss may result from failure. Game over.