Research: Industrial Networks Are Vulnerable To Devastating Cyberattacks 76
Patrick O'Neill writes: New research into Industrial Ethernet Switches reveals a wide host of vulnerabilities that leave critical infrastructure facilities open to attackers. Many of the vulnerabilities reveal fundamental weaknesses: Widespread use of default passwords, hardcoded encryption keys, a lack of proper authentication for firmware updates, a lack of encrypted connections, and more. Combined with a lack of network monitoring, researchers say the situation showcases "a massive lack of security awareness in the industrial control systems community."
oh my! (Score:1)
Re: (Score:1)
I'm pretty sure this exact report was created 10 years ago, and has been republished every year since.
Is it just a way to steal more federal funding for these "studies"?
The install base is screwed, and nobody wants to spend any money to improve the situation. Profits over everything else.
Why didn't someone tell us this earlier? (Score:2, Informative)
-- Iran nuclear program
Re: (Score:2)
Re: (Score:2)
It is very old news. And even when it was news, it was no surprise to experts. People need to see a catastrophe close enough that they get scared in order to invest into security. Because if they do not, they can get rich faster, or so they think...
Industrial network (Score:3, Interesting)
Re: (Score:3)
Follow the money.. Who pays the bills? Do what they say...
Seriously, keeping your factory's networks separate is a pretty simple firewall issue for someone competent to install and configure it. I'm not sure how this cell phone connection is going to work, but there ARE ways to make cell phones connect to you via VPN's that can be made to require usernames/passwords (not to mention specific devices) before you are allowed to connect. There are solutions out there to do what they ask, they just cost a li
Re: (Score:2)
So my "follow the money" joke, really should be this. IF the people in charge are asking for it, find and suggest a solution that can do it safely. If they are not willing to pay for your solution, find another, albeit less safe solution and present it with a list of assumed risks. Rinse and repeat until you have a solution they are willing to pay for with risks they are accepting, then do that.
They want easy, and cheap. That limits you slightly...
Re: (Score:2)
Re: (Score:2)
So my "follow the money" joke, really should be this. IF the people in charge are asking for it, find and suggest a solution that can do it safely. If they are not willing to pay for your solution, find another, albeit less safe solution and present it with a list of assumed risks. Rinse and repeat until you have a solution they are willing to pay for with risks they are accepting, then do that.
They want easy, and cheap. That limits you slightly...
Understood, but the POINT of this little exercise is to educate the people asking about the true costs and risks they are assuming before you implement anything. Responsible bosses appreciate this kind of iterative process that allows them to choose their level of risk and cost, and if you are dealing with irresponsible bosses who will come back and blame you for failure at a later date, you are going to need the CYA documentation that shows THEY agreed to this and you already told them of the risk.
If the
Re: (Score:3)
I'm with you so far.
In my experience, any "solution" that you present will be understood to do everything that they wanted.
Even if you say that they cannot have X at $Y
Re: (Score:1)
If you have a management that is fighting you for 2-4 thousand dollars of work at most, you have bigger problems to worry about.
Re: (Score:2)
If that's true, and many times it is, you are playing a loosing hand to start with. You work for people who don't care about you or value what you say and do and it's a BAD place for you.
Best thing to do if you work for people like this is start to plan your departure. Update your resume, get your online profiles updated and start looking at the employment ads. It might be a very good idea to start putting money away for the "rainy day" that's surely coming.
Maybe YOU can be that consultant they hire over
Re: (Score:3)
Re: (Score:2)
That's why you present your "solution" before implementation and that includes documentation of the provisos and risks they are taking.
It may not save your job when the chips are down, but having a bit of hard documentation that you told them what the risks where and they choose not to spend the money to eliminate that risk is always a good thing. Besides, if they assume the risks, then fire the underling when a risk bites them, you really don't want to work for them anyway because it's just a matter of
Re: (Score:2)
Re: (Score:2)
Keep in mind there is a major difference between monitoring and controlling. To control/run you have to be able to provide input into the system. It is this input access which opens the door.
Obligatory "why" post (Score:5, Insightful)
Re: (Score:3)
Re: (Score:2)
See comment on application level gateway. I work at a plant where we have access remotely, but no our control system is not connected to the internet. There's layers of VPN, firewalls, and even at the lowest level the final application is a single program served up via citrix.
Re: (Score:1)
Re: (Score:2)
Re: (Score:1)
giving you strata 5, plus or minus 10 seconds. duh. may not be good enough.
First of all it's "stratum" not "strata".
Second, adding one layer between add's exactly one stratum, no less and no more.
Most public time sources (GPS/PPS/...) are between 1-3 and thus claim jumping to stratum 10 with single firewall, let's call it bollocks.
Here's a good short explanation how it works. [endruntechnologies.com]
anyways these days one can build a strata-1 NTP server from GPS and run it on the internal network.
Building just one doesn't make sense if you take that option. Rule of thumb is that whenever you need local NTP servers, you also need to have 2-3 to provide you enough redundancy and make other important arr
Re: (Score:2)
Corection, kerberos & nfs difference between client and server generally should not be more than 5 seconds, so above should be +/-2.5 second.
That's a protocol design bug.
Specifically, there's actually no reason that protocol traffic wouldn't include a "this is my idea of the current time" in the requests and responses so that delta times could be locally calculated from the packet contents on the receiving end. This would work, no problem, for a protocol like NFS.
Kerberos is more of an issue, but since all parties have to trust the ticket granting system as the trusted third party -- so you might a well trust their timestamp as well, since you'
Re: (Score:1)
Re: (Score:2)
Why? Just use a GPS disciplined NTP server locally.
Re: (Score:2)
In the past low skilled staff would have to be in place, drive to or be on site 24/7.
The cost savings add up for the brand but the quality of the network installed expected correct commands on a private network not a network open to the world.
Years later all the limited networks open to the "net" per nation have been transversed and studied by a long list of people and other nations.
The
Who knew? (Score:2)
It's not "industrial," ... (Score:2)
... it's everything.
Security will continue to be a low priority until we assign blame and litigate.
Re: (Score:2)
Re: (Score:2)
I'm in the fucking business, so punt.
Robotic Surgeons? (Score:3)
Does it make anyone else uncomfortable that this story about industrial networks being vulnerable to cyberattacks follows immediately after a story about robotic surgeons?
Re: (Score:2)
Re: (Score:2)
Not really. If you look at the likelihood of being in surgery when the network goes down, or the surgeon gets hacked, it's pretty much negligible.
Not for that one unlucky guy... Someone will be there.
Letting everyone know won't help. (Score:1)
This is why Cloudflare got four of its routers wiped out during that last October DDoS. As soon as the network infrastructure was known and exploits located, it was the attack point. Security failure.
And this is only going to get worse.
obvious solution: (Score:3)
look, none of this is a problem as long as nobody asks about the worst case scenarios.
Re: (Score:2)
look, none of this is a problem as long as nobody asks about the worst case scenarios.
Umm... you mean what's the worst that can happen?
dammit man, haven't you ever seen a disaster movie?!
Re: (Score:2)
In a worst case scenario, this would make all politicians of all parties at least minimally competent and decent human beings (again, at least minimally), would solve the national debt and at least one major societal issue...
(waiting)....
(waiting).......
(waiting)...........
Oh well, it was worth a shot.
Different approach (Score:2)
Re: (Score:2)
There is this piece of Cat 5 that isn't remotely hackable. Unless it's tapped, or if someone puts an inductor on it, or if they use TDR to estimate the length of the wire to figure out the distance between routers and discover where the Intrusion and Detection Systems are located.
Re: (Score:2)
But that wont get a cyber security contract long term to "fix" the system after every expensive logged intrusion.
The new networks have one good plus, wealth creation for the support, upgrade aspect.
Air gaps don't have backdoors (Score:1)
"Backdoors also exist in the form of hidden accounts originally created for maintenance that can provide cover for attackers. In particularly insecure facilities, antiquated and unencrypted connections to the Internet that allow engineers remote access to their networks act as pathways an attacker anywhere in the world can take toward the network in her crosshairs."
Air gaps as much as you can, air gaps between you control networks, and your mail/office/facebook network. Don't trust vendor kit, its riddled w
Re: (Score:2)
If there had been cyberattacks in the early days (Score:1)
When networking of smart devices was still on a relatively small scale, a cyberattack wouldn't have done much harm, but afterwards, manufacturers, and more importantly, their customers, might have wised up. Stuxnet was a warning, and I think it has to some extent been heeded, but already by then the existing infrastructure was so vast that a major overhaul would have required a commitment and leadership that isn't there.
Re: (Score:1)
Good luck fixing it too. (Score:1)
I work in a small manufacturing company, all it would take is one malicious person to get on the network, send some specially crafted EIP packets to some of our PLC's and production is fucked. I keep saying we need to segment and isolate the industrial network from our poorly managed corporate network, but it gets ignored because "3000 is a lot of money to spend on some computer stuff."
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
Re: (Score:1)
We're not showing lack of awareness... (Score:2)
It's not a case of lack of awareness, it's a case of mostly not giving a shit. We don't use most of the encryption features or hardening available between control systems on our site either because quite frankly we don't expect to and we don't need to. Actually I was quite critical at the last Schneider conference where they were talking about the encryption they are adding allowing you to connect multiple SCADA systems together directly via the internet. My comment to the presenter was "Why should I care a
New research shows massive lack of security .. (Score:2)
New research - new research ? - this has been known about for at least a decade ref [huffingtonpost.com]. and the solution is, don't connect your switches directly to the Internet, connect them through VPN [techterms.com]s running on embedded hardware.
So how about... (Score:2)
Are Industrial Networks also Vulnerable To non-devastating Cyberattacks?
How is this news? (Score:1)
This has been true for at least 20 years.
The industrial controls industry is the most backward corner of the tech world, inhabited by an old guard that mostly doesn't even understand networking, let alone security. The newer recruits generally come from an EE background, so they also generally have no knowledge of how to secure critical infrastructure. Most started in the era where inter-device/machine communication was via serial and all these systems were simple air-gapped (not for security, but because t
It's a very fine line (Score:2)