Slashdot Mirror
Israeli Security Company Builds "Unhackable" Version of Windows
New submitter Neavey writes: Sounds too good to be true, but Morphisec, an Israeli startup, claims to have built an unhackable version of Windows. Its not yet publicly available, a red flag if ever I saw one, but internal testing has had a 100% success rate: "In a statement for BI, Dudu Mimran, the co-founder of the company, describes this new OS version as the Windows that 'Microsoft should be doing,' explaining that, while the platform was initially designed for government use, it can be actually installed by any enterprise that wants to make sure that no hack is possible.
Basically, this operating can block any zero-day attack, the founder says, thanks to the operating system randomizing all memory, which means that the hacker cannot target the computer memory and compromise the data stored on the drives."
What things memory randomization does not fix, left as an exercise for the reader.
I hope everyone at that company is prepared for a long week.
SJW's don't eliminate discrimination. They just expropriate it for themselves.
It is being offered to the mullahs on a flashkey.
You may want to take a look at some of this company's other products, including flying serum and invisibility powder.
Just remove all input and output capabilities, and the power supply. Most secure computer in the world.
According to my own internal testing, of which i've done none.
Per the article, they've raised money and it's under development. Sounds more like they're at the generate some buzz for some more money stage of development.
But I concede that randomizing memory (contents) does make a system pretty secure.
This company (or whoever wrote TFS/TFA about them) seems not to understand the concept of a zero-day vulnerability.
It is ridiculous to say that one is not vulnerable to zero-day attacks. They are, in security parlance, the "unknown unknowns" - the things you don't even conceptually know of as vulnerabilities right now. One cannot design a networked computer system with any functionality whatsoever in which they can somehow know and anticipate the "unknown unknowns" (as opposed to the known unknowns, some of which can be mitigated if you're lucky).
The unknown unknowns are, by definition, *not yet known*, so you can't design a mitigation against them until *after* you are aware of them. If awareness comes in the form of a zero-day hack, then you will fail to defend against the attack at the time it hit due to your lack of information about the attack vector.
Also, unless this company has full access to all Windows source code for the build they have, it is very likely that one singular memory-based mitigation will not be effective against every possible attack vector that exists in the Windows codebase. So unless they have performed full formal methods verification of the entire Windows codebase to guarantee that there are no "unknown unknowns", and then fixed every security vulnerability that exists in the product in the original state in which they received it from Microsoft, this is basically snakeoil.
Also, don't we already have ASLR? The mind boggles at the stupidity of these people. Who do they seriously think is going to buy this?
Actually, forget I asked. They said their target was governments. I have no doubt they will sell thousands of licenses.
has had address space randomization for how many years? Hardly unexploitable still...
Oh yeah, I've seen builds that were 100% solid on internal testing. Not a thing wrong with it according to automated tests, scripted manual testing, smoke testing, and random usage testing. Not a thing! A million monkeys could bang on keyboards all day long and nothing would break. Much simpler programs than an entire OS, mind you. But still, they were bullet-proof, air-tight, divine works of software engineering.
Then we pushed them to production. Murphy's law is a moooootherfucker.
Captcha: enraging
...for approximately 15 minutes to hack the unhackable today and then resumed normal business with smirking faces all around...
Loading...
I mean, if it's invincible to tech-based hacks, kudos to them... but the other side of that is the wall of gullible idiots that will be manning the "unhackable" systems. Some quick social engineering and their impenetrable fortress will have more holes in it than Swiss cheese.
Memory randomization has been around a very very very long time. It's not going to help with logical programming errors.
1) Disable all network access. 2) Disable all external storage access (USB, DVD, etc). 3) Most importantly, disable all user logins.
Why do people still claim these things, and why to techies (not marketing people) consent to attaching their names to such nonsense?
Stupid because:
1) No, it is not unhackable. Throw a contest with a bounty to easily prove this.
2) 99% of "hacks" work through social engineering nowadays, and these work regardless of how secure your software is.
3) Selling your own modified version of Windows will get you sued by Microsoft very quickly.
Slashdot has often featured articles from Israeli companies that seem to me to be fraudulent. For example, The Car That Makes Its Own Fuel. That Slashdot story links to this article: The Car That Makes Its Own Fuel.
Easy. You don't need to worry about upstream updates 'cause the system is unhackable.
Duh. Idiot.
Are they just talking about Address Space Layout Randomization? Let's see - Wikipedia says [https://en.wikipedia.org/wiki/Address_space_layout_randomization] for Windows - to turn it on edit a registry key. Is that what this company did, "create" a version of windows with a registry key set?
... but what are the chances of that?
Security relies on certain assumptions.
If I have a military base, I assume that whomever comes to attack my base has fewer guys with guns than I do... and I generally it will be a cold day in hell before they'll get very far into the base.
And you assume other things... you assume that your security people can tell the difference between someone with security clearance and a birthday clown.
We assume that the people with clearance obtained it legitimately.
We assume that the people that were given security didn't subsequently decide to sell us out for hookers and blow.
Assumptions.
And there are good assumptions... assumptions that really will hold under most circumstances and bad assumptions.
And good security is basically a process of separating out good assumptions from dumb ones. Then recognizing that your dumb assumptions were a convenient fig leaf you put over serious vulnerabilities that you actually don't have a good solution for...
And then you need to actually come up with a GOOD assumption that covers for what were previously laughable assumptions.
If your security is based on interlocking layers of good assumptions... are you unhackable? I don't know... its a question of perfection and perfection is hard in this universe. BUT... really fucking good security? Near perfect? Sure. I mean... you can do "excellent"... excellent is possible.
But that's not to say that even good security should be discounted as crap. Good is often the best security possible because excellent requires time and money and competent management and users that don't have their heads wedged up their asses.
Now will good security keep ze germans out or whatever? Typically yeah. Even good security is a bitch to get through even for a state sponsored hacking team.
What keeps embarressing people is SHIT security or NO security.
That is what keeps failing. Not "good security"... not "excellent security"... not "perfect security"...
F'ing none at all keeps failing.
So... lets not geek out on the "perfect" or "unhackable" claim. And instead lets focus on whether or not the change to the OS makes Windows have "good security". If it accomplishes so much as that then we're doing well. If they pushed it up a notch and it's EXCELLENT... Then we're doing very very well indeed.
I've decided to stop wasting my time responding to AC trolls/sockpuppets... so if you want a response from me... login.
The headline is crap, of course.
That said, it's not too hard to have a version such that you know it's unaltered when you boot each morning. You do basically a live CD, booting from a read-only lun.
Just as you separate a normal user USING the machine from an administrator account UPDATING the OS, you can have the OS basically read-only during use and set it to writeable only when you need to update the software. That change is done outside of the OS, either via the NAS or the hypervisor.
In that way, you can come in eqch morning knowing your Windows system hasn't been hacked (past tense). As soon as you open IE, though, you could get a new exploit. That exploit disappears when you shut the machine down, though.
Everything was going very well, until Shlomo installed Flash player.
I think if Windows ran everything in something like a sandbox, where programs couldn't communicate with programs outside itself, and saw its own version of a disk system which only had itself on it, things wouldn't be bad for starters. A virus then couldn't then spread to other files on your filesystems because each program couldn't access programs outside itself.
.exe, I would try out a lot more software.
It doesn't help much for legacy software, but a special memory section could be used for shared memory, and a special disk location could be used for shared files.
A system prompt would be needed before installing driver files or changing things on startup.
This doesn't stop a keylogger from getting you though. There are ways of stopping keyloggers, but no need to get to complex stuff when people will want to shoot holes through my theory "Windows as a filesystem sandbox mode". I think about this a lot since it doesn't seem like several OSes are designed to operate in the Internet environment without getting hosed by running the wrong file. If Windows could be secure from running an occasional malware
God spoke to me
Which then poses the question... just how is this any different, let alone superior to Linux's PaX patchset - which offers ASLR since 2000 - or even grsecurity?
My Commodore 64.
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
100% Secure = 100% Unusable
Security is a balancing act between usability, functionality, and safety.
You'll never get 100% in any of those without having less than that in the other two categories.
Sure, they may get closer to 100%, but at what cost? Is the machine running slower? Does it eat up huge amounts of HD? Does it take a 5 minutes to verify an authorized users biometrics before allowing them to do anything and if they leave it's immediate 'secure' area it totally resets?
Not that those are what this one is or isn't doing, I was just illustrating the point that you can't have perfect security, and have a usable machine because there are always trade- offs. Especially since it's under the rule of diminishing returns. Although one great way to easily improve security is to remove humans from the loop. Of course, then you are just talking about some kind of backend or infrastructure type thing since it's only 'users' would be other machines, and even that can be compromised by compromising the machines that are allowed to be users.
That's why I say that a machine that is totally secure, is also totally unusable. It's the only way to prevent the machine being compromised, but that's not really any good to anyone either.
You are correct, nobody ever hacked the Titanic.
BIOS is dead. With EFI, most of the boot code is in the efi partition, on the "disk" which is read-only courtesy of your san, hypervisor, or the fact that it's a cd-rom.
There is a limited firmware on the motherboard which loads the initial efi file. That could, in theory, be compromised, except that if you virtualize, you could also set that read-only in the hypervisor. So your virtual machine pretty darn safe. The host machine needs to be secured , but it doesn't need an operating system, just a hypervisor. That's quite a bit safer than running a full desktop OS.
The headline is crap, of course.
That said, it's not too hard to have a version such that you know it's unaltered when you boot each morning. You do basically a live CD, booting from a read-only lun.
Just as you separate a normal user USING the machine from an administrator account UPDATING the OS, you can have the OS basically read-only during use and set it to writeable only when you need to update the software. That change is done outside of the OS, either via the NAS or the hypervisor.
In that way, you can come in eqch morning knowing your Windows system hasn't been hacked (past tense). As soon as you open IE, though, you could get a new exploit. That exploit disappears when you shut the machine down, though.
Or you can put Deep Freeze on it and have the same thing every time you reboot, morning, noon, or night. MEOW!
So, I hope they aren't trying to patent too much of this idea. It's been prior art for 10 years. Here is a link to an archived version of my post: http://www.derkeiler.com/Newsg.... It is all I could find from my phone.
I don't mind them using the idea. I posted it publicly hoping someone would. But they can't claim to own the idea or prevent others from using it.