Releasing rainbow tables to accelerate Net-NTLMv1 protocol deprecation | Hacker News Reader

Releasing rainbow tables to accelerate Net-NTLMv1 protocol deprecation

cloud.google.com

162 pointslinolevan89 comments

I recall using ntlm rainbow tables to crack windows hashes in high school in like 2008?

Amazing that this is still around and causing someone enough of a headache to justify spending money on.

Also amazing what a teenager with lots of free time and a bootable Linux usb can get up to.

LM, nthash aka NTLM, net-ntlmv1 aka ntlmv1, net-ntlmv2 aka NTLMv2. Challenge response stuff is different. Naming here is painful.

Ah Microsoft and naming things... Name a better combo

But fair enough, I don't recall which exact version I was mucking with that long ago.

Ya they just announced they are renaming security algos to copilot!!! story here -> https://dubious-adware-breach-scam@is.gd/WVZvnI?exploit.bat

Love this. Classic microsoft.

A few years ago i was doing some vm things in azure. Hadnt touched azure before, and spent 10+ minutes of frustration trying to figure out how to get amd64/x86_64 things started, as the only thing i could find was "Azure ARM", and on googling, "arm" here means azure resource manager... ARGH why does microsoft insist on using existing names and acronyms!?!?

Because in their eyes if something was not invented here, it may as well not exist :-) they haven’t managed to cure this sickness in decades.

I was part of a user study on Azure back when it first rolled out-- they were looking for seniors with an AWS background to participate in UX research, and I remember walking out of that study with imposter syndrome for the very first time. Spent 60 minutes totally unable to do the thing I wanted to do when I was introduced to Azure for the first time, and I remember thinking... am I a fraud?

No! Not this time, at least. In hindsight everything was named and organized terribly and it hasn't improved much since.

net-ntlmv1 rainbow tables have been around forever too though, the same attack documented in this blog post has been hosted as a web service at https://crack.sh/netntlm/ for 10+ years

Yeah, but now it's Google! Google!

There used to be a joint online project to compute these tables in a SETI like distributed system. Everyone who contributed their CPU cycles, could use the tables. And yeah, around 2005-2008.

yep, that and also can use cain and abel even back then... hardest part was putting whatever network card in promiscious mode.

Yes!! That was the software, thanks for the memory trigger

To be vulnerable to this, what sort of dumb things are end users doing?

I couldn't immediately figure out here whether we're talking

0. Microsoft's supported products default enable this worthless "authentication" feature

1. Microsoft's supported products provide such a feature behind a UI that's not clearly marked "Danger: Do not stare into laser with remaining eye"

2: Microsoft does still support this, behind some Registry nonsense most users do not understand and once enabled it doesn't turn on the "I am a toxic waste dump, leave by nearest exit" warning signs on affected machines

3: Microsoft doesn't support this at all but some 3rd party commercial stuff does and customers really do love their crusty archaic 3rd party garbage

4: But this long abandoned SCO machine we've kept on life support for twenty years!

5: What does "supported" mean? Windows NT is scary, we're still on Windows 98 here.

This is mostly an issue in active directory networks. Usually the reason people give as to why they still have this enabled is due to some legacy system that can't authenticate via Kerberos or at least NetNTLMv2. Worst case is if they then enable NetNTLMv1 on the domain controllers, even if the DC acts as a client. Using authentication coercion, this is a pretty quick win for an adversary.

Ok, so it's a 2 on my list.

Microsoft needs to make this forcibly change the UI so that users can see "Oh! I'm using crap low security Windows". That lights a fire under people to actually get it fixed.

Using any variant of NTLM is insecure, which is why Microsoft is phasing it out in Windows 11/Server 2025. Which means we should be free of it some time around 2060.

You have to actively go out of your way to have this enabled. But large or older companies always have some old machine (can be as dumb a an old but very expensive printer) that isn't updated.

Even today the only reason to use samba 2 in 90% of companies where it's enabled are old appliances.

At some point device X isn't working, employees complain, IT say they need to buy a new very expensive replacement and after much argumentation they come to the agreement to enable that legacy horror support until the purchase can be made. Which is then never made.

For those interested: The SHA512 file lists 4096 files. Each file is 2 GiB. That means 8 TiB (or about 8.6 TB) of storage required.

Rule of thumb: Saturating 100Mbps moves roughly 1TB/day.

Really curious how this was discussed with the legal team...

"We're releasing hacking tools to allow others to break into poorly secured computer systems... But we are doing it with good intentions so it won't be illegal right??"

Isn’t this more or less a case of “illegal numbers”?

Yes. This is no more pernicious than releasing a multiplication table.

"Educational use only", just like every other skiddie tool you can freely grab on the Internet these days.

(The bane of my existence thanks to everyone and their mother releasing ESP32 Wi-Fi/Bluetooth/etc deauther 'tools'...)

Maybe actually secure your network then :D

802.11w exists

Doesn't protect against an attacker who just sends RTS + CTS packets every 50 microseconds, preventing all other wifi devices from transmitting anything (inc beacons).

As a bonus, even wireshark won't detect anything for debugging, because the RTS+CTS handshake typically can't be sent to the linux kernel+higher layers.

An ESP32 is capable of doing that on all wifi channels 'simultaneously' (ie. round robin, but getting back to the start channel within the timeout), effectively blocking all wifi.

Second this.

I didn't expect Google (Mandiant) to release rainbow tables ever. Curious what changed internally to make that acceptable now.

This empowers script kiddies, but not significantly moreso than they already were. Of all the places this is still in use, they've been exposed for years, so this isn't likely to result in a a bunch of new exploitations.

However, it's most likely to be used by governments, with legacy servers that are finicky, with filesharing set up that's impacted other computers configured for compatibility, or legacy ancient network gear or printers.

I wonder who they're pushing around, and what the motivation is?

Mandiant is Google's incident response consulting business. Having worked for many years in that field myself (though not for Mandiant), they're probably sick of going to the same old engagements where companies have been getting owned the same way over and over again for the last 15 years.

What releases like this do is give IT ops people the ammunition they need to convince their leadership to actually spend some money on fixing systemic security problems.

> Mandiant is Google's incident response consulting business

Consulting business? I was under the impression (from Google Reader) that if users aren’t in the millions, then they’ll kill the project. How could they also run a high-touch consultancy?!

> they're probably sick of going to the same old engagements

Hmm… consultancies love this type of recurring revenue - it’s easy money

> Consulting business? I was under the impression (from Google Reader) that if users aren’t in the millions, then they’ll kill the project. How could they also run a high-touch consultancy?!

Google also has the Project Zero which doesn't fit into Google business culture either. I wonder if Mandiant is paying for their payroll.

Project Zero has been around for 8 years before the Mandiant acquisition.

My bad. Still not sure which business unit is paying for their payroll.

Google is a quarter million person company (if you count full time, temps, vendors and contractors).

Google Cloud is basically an entirely different company than Search or Maps. Cloud will happily sell you $10m in compute a year and a value add $400k of security consulting.

I suspect Mandiant hears a lot of "this is impractical to exploit so we don't care" from their clients. Now they have a compelling rebuttal to that.

It also empowers IT depts and cybersecurity people to be able to easily build a PoC to show why moving on from the deprecated protocol is important. In many white-hat jobs you can't just grab rainbow tables from a torrent, so a resource like this is helpful. For the grays and black hats, they've had access to rainbow tables like this for a very long time, so no change there.

Out of curiosity, why can't white hats grab rainbow tables from torrents? Is it about seeding?

Its less about torrents being the delivery mechanism and more about bringing data from a potentially unknown source, under potentially unknown licensing, and distributed for a potentially unknown reason into the corporate computing environment.

Torrents would be a perfectly valid way for Google to distribute this dataset, but the key difference would be that Google is providing it for this purpose and presumably didn't do anything underhanded to collect or generate it, and tells you explicitly how you're allowed to use it via the license.

That sort of legal and compliance homework is good practice for any business to some extent (don't use random p2p discoveries for sensitive business purposes), but is probably critical to remain employed in the sorts of giant enterprises where an internal security engineer needs to build a compelling case for spending money to upgrade an outdated protocol.

Any business that needs convincing to move on from anything labeled NTLM does not care what "nerds" have to say. They are either one of those "I'm not spending money on something that works" or stuck with such legacy technical debt that at this point, removing it from environment is too costly to even consider so executives kick it down the road.

Or it’ll be like a conversation I had yesterday, where the “Active Directory guy” who’s been in the job for 20 years doesn’t even know that there’s an NTLMv1 and an an NTLMv2.

You've been able to find these for years. In fact it's entirely possible they just grabbed some or all of them out of an existing torrent originally.

It would completely not surprise me if there are automagic attacks on net-ntlmv1 at this point against some cloud hosted storage. This has been doable by anyone since like 2016 if you had the space and weren't prevented from using that protocol version.

Mandiant releases rainbow tables for a 25 year old broken protocol because enterprises still won't disable it. It seems like sometimes the best security tool is just making the risk impossible to ignore.

Yeah that protocol is very very broken. I recently did an ntlm plugin implementation for Caido [1] and I had to fork our crypto JS module to add back MD4 and 3DES.

[1] https://github.com/caido-community/ntlm

Didn't l0phtcrack do this like 25 years ago?

I actually got a job that long ago by using l0phtcrack to expose an admin password for an NT4 network.

NTLM is not Net-NTLM- l0pht did ntlm

And terrorism is just an abstract way of securing underprepared government facilities.

if you think about it, the united states terrorism risk insurance act distorts the terrorism market by subsidizing terrorism-generating activity

https://home.treasury.gov/policy-issues/financial-markets-fi...

Tangent: this link looks like a right-wing propaganda site; I wouldn’t trust any information there without verifying it with trusted sources.

> Thanks to the President’s decisive leadership in the face of radical left-wing obstructionism, the Department of the Treasury has now resumed normal operations.

its an official government website

This is like reminding that there are CVSes from 2010. Yes there are. And there are plenty of vulnerable systems.

They decided to not fix the vulns (either directly by not patching, or indirectly by not investing in cybersecurity). So exploiting them is somehow an act of mercy. They may not know they have a problem and they have an opportunity to learn.

Let's just hope they will have white or gray-ish hats teaching the lesson

They're just dumping them out as 2GB blobs onto a cloud? Where is the zippy search UI? Very lazy behavior for the hyper giant Google.

Right? I feel like rainbow tables for NTLM have been around for decades, though at-cost. This seems incredibly low effort on Google's part.

They don't know how to count that low.

https://www.lesswrong.com/posts/koGbEwgbfst2wCbzG/i-don-t-kn...

Why would you want a search UI for a rainbow table? That makes no sense.

That's the most likely use case for anyone. You want the one-offs when cracking, and not a compressed 2GB blob of data that may or may not have your answer.

B. The king of search has held the goal of organizing and making information useful and searchable since dinosaurs roamed the WWW.

https://www.google.com/intl/en_us/search/howsearchworks/our-...

C. It's just lazy and shows that they don't actually care anymore. Making a custom search has been their bread and butter for decades. The last step is the easiest but they could not be bothered.

Holy smoke. I honestly thought the 90s called and wanted their Windows exploits back (TFA mentions 1999). I do remember talk about this from many moons ago.

But we are in two-thousand-twenty-FUCKING-six.

It's unbelievable. Just plain unbelievable.

pretty cool

Can't wait for someone to decide one of protocols used by google needs to be deprecated.

Plenty of protocols used by google over the years have been deprecated. The difference being that google actually stops using insecure protocols when they are discovered to be insecure instead of trying to sweep things under the rug.

Keep in mind we are talking about a protocol from 1987. How many protocols from 1987 is google currently using?

Google does whatever is convenient and makes them money. Altruism was never part of the equation.

Sure. Not being hacked is good for business.

Keep in mind that google is primarily a cloud business. That means that they take on a lot more of a risk, as when they are hacked its a them problem vs traditional software where its much more the customer's problem. Security is very much about incentives, and the incentives line up better for google to do the right thing.

It's more about when Google assumed full control of the cloud, the browser, the OS, and everything in between they self-appointed themselves as the unelected standards board of the Internet, and forced everyone else to follow their whims and timelines. Some of which are completely insane.

What are the policies you view as "completely insane"? I have some I disagree with like how they've handled things like Manifest v3 in the browsers, however there are still alternatives like Firefox anyway. However I think in terms of web standards some of the things they have pushed are also helpful. It's been much nicer having a much more consistent web browsing experience with less things like "You must use Internet Explorer on this site".

I feel like web browser and website standards are one of the main areas Google has a lot more control of policies. Is there somewhere else they have much control of for standards?

Re: the IE thing. Only Apple's insistence on Safari in iOS is stopping sites from basically being chrome only.

>It's been much nicer having a much more consistent web browsing experience with less things like "You must use Internet Explorer on this site".

What browser do you use?

Because I've definitely run into this but s/ie/chrome/ but with no helpful message. You just have to guess that that's why it's broken

And you‘re implying making Windows networks less insecure is completely insane?

This is such a negative reading of the situation. You’re talking about something that has been compromised for TWO DECADES.

At least now nobody can pretend.

I for one hope that this hastens the demise of every remaining use.

Well, you'll be waiting 20 years or so post-deprecation if you want an equivalent timeline.

Google thrives on being the Internet's biggest bully.

It turns out when nerds get a billion dollars they like being bullies too.

Google does that every Tuesday

> under 12 hours using consumer hardware costing less than $600 USD

Great, so someone with half a motherboard can break this hash

Or 1gb of ram, but not both

"To demonstrate how crappy most front door locks are, to boost our company's social media cred we will be leaving drills and a dish of bump keys at the entrance of the neighborhood."

NTLMv1 rainbow tables have been available for 15-20 years. The only thing new is that Google are publishing theirs.

NTLM is often used for more of the underlying technologies, some more secure than others… nthash, net-ntlmv1, net-ntlmv2. There’s a little more complexity here and this is different than the stuff that was out 15 years ago

> this is different than the stuff that was out 15 years ago

This stuff was out at least 10-15 years ago. It’s different from the ancient local ntlm hash cracking everyone used to get admin in high school, yes, but it’s not a novel technique.

on cursory google, https://github.com/NotMedic/NetNTLMtoSilverTicket/blob/maste... is 6 years old and was old news when it was committed, and https://crack.sh/netntlm/ has been around online for at least 10 and I think more like 15+ years.

Microsoft has deprecated NTLM and is actively ripping it out of windows.

https://support.microsoft.com/en-us/topic/upcoming-changes-t...

Windows 11 is probably the last version that will contain NTLM (and hopefully NTLMv2). Going forward everything will be Kerberos or Oauth based.

Ironically enough, the things that tend to break first when you try to turn off NTLM are still Microsoft products like ADCS.

You're not wrong, I just want to point out this is net-lmvm1, which is different and more complex. Not functionally meaningfully more complex to an adversary with a few hundred USD (almost typed LSD) in monies. But technically larger tables. That being said I'm in agreement that this has been known problem for 10+ years, and Google is just saying the horses are so long out of the barn their grandchildren are grazing.

you say that like it's a negative analogy

The bad guys already know you live in a bad neighborhood and have been closing your front door with a plastic combination lock you got in a Happy Meal 40 years ago. They can already come and go at a whim. This is Google letting you know that your crappy lock is pre-broken to encourage you to upgrade to literally anything else.

It's certainly morally and legally dubious to facilitate attacks on things that others choose to use in within their own private domains, just because you disagree with that choice. But that's how these people roll.

It's been 15 years since this was known broken. If you had children when it was not known broken, they'd be almost old enough to drive in most western nations.

At some point the line must be drawn.

Some are very entitled to drawing lines on someone else's property. Why don't you mind your own business?

I mean this kindly, but if you're still using net-netlmv1 on anything that matters, you need to pay much more mind to your own business because even the original vendor of it has been telling you to get off that since 1999 because it is not safe.

If you're using it on something that doesn't matter, then it also doesn't matter that rainbow tables any attacker could have already had for a decade are slightly more available.

I wonder how the Mandiant acquisition is regarded within google.

Was it a success? Is Mandiant a cash cow or was it basically an acquihire?

The big "contact mandiant" button next to the post feels a bit like trying to stay relevant and acquire more customers.

> trying to stay relevant and acquire more customers

Is there any business that does NOT try to do this? Why wouldn't they?

Funeral service providers, for obvious reasons.

Reply on news.ycombinator.com