Researchers Reveal Secrets of SHA-1 Hash Collision
LAS VEGAS—Elie Bursztein, Google's pb anti-fraud researcher, began his talk hither at Blackness Lid 2022 with an understatement: "Information technology has been a long and interesting journey over the last few years." In fact, Bursztein has spent the better part of 2 years working to create an SHA-1 hash collision, which means he demonstrated non but how he could practise the impossible, just how information technology'southward done quickly and efficiently.
What'southward a Hash?
There are two critical points about cryptographic hash functions like SHA-ane. Get-go, it's a one-style process. Yous tin't accept a hash of a file and reconstruct that file. More chiefly is that no two hashes are ever the same. "If y'all were to hash ii files, even if they have one bit of departure, the hash will be different," explained Bursztein.
And while you can't fume a cryptographic hash (non fifty-fifty in Las Vegas), you can find them everywhere. Cryptographic hashes are used to verify the legitimacy of digital contracts, since you tin compare hashes and confirm that document hasn't been changed. Hash functions are used to digitally sign HTTPS certificates, which verify a website'southward identity and secure your connection. These functions are as well used to digitally sign software, and then your computer can tell the difference between a legitimate software update from Microsoft and a bogus i designed to have over your computer.
You lot've probably heard most hashes in the context of data breaches, likewise. When a company announces that its data has been stolen, you'll typically hear that passwords and other information are secure because they've been hashed. This can mean a lot of things, but hash functions are great for passwords. Instead of storing the password, a company simply retains its hash. When you enter your password, it's hashed and compared to the hash the visitor has stored. If it'south a match, you lot're in! And if it's stolen it doesn't matter because an aggressor can't utilise the hash to obtain your password.
When Hashes Collide
A hash collision is when two different files end upwards with the aforementioned hash. The benefits are obvious; if you have a phoney contract only information technology has the same hash every bit the original contract, you can sneakily demand any terms you want.
Simply like all things in cryptography, a hash standoff is non easy. "You cannot employ beast strength," Bursztein said, referring to the practice of merely throwing tons of computing power at a cryptographic problem and hoping to find the answer. "Nosotros estimated that if you lot wanted to creature force this, information technology would take 12 one thousand thousand years for 1 collision."
The massive time and calculating power required to solve the equations that make upward cryptographic tools is what makes them work. It might be possible to notice the reply, merely non within the lifetime of the attacker or whatever of their descendents.
I am no expert in the advanced mathematics that went into the work of Bursztein and his collaborators. Even in the talk, Bursztein breezed over some subjects. "There's 15 years of cryptanalysis in hither," he joked.
Part of Bursztein's approach involved solving individual steps in the 80-step SHA-1 hash procedure. His team solved the offset xvi steps, and found that the next eight steps were extremely predictable. After that, he admitted, his team hadn't cracked the procedure. Merely importantly, they didn't need to.
"We have been able to push button back the complexity of the function," he said, with the remaining steps being within the grasp of computing power.
Bursztein's talk ofttimes veered into highly technical territory. While it's not unusual for Black Chapeau talks to involve code that looks bulletproof to the layman, this had to practise with the bodily structure of files, where data is located, how the files are constructed, and how pieces of the file are addressed by the hash part. Information technology'southward like playing with the atomic structure of fe to turn information technology to aureate, and the result was, to my untrained eyes, every bit magical.
The results were impressive. Instead of taking 12,000,000 GPUs operating at peak capacity for ane twelvemonth to find one standoff, he found he was able to find a collision using but 110 GPUs for one year. More in-depth data, as well every bit a tool to brand your very ain hash collisions, is available at the project's website shattered.io.
But Who Cares?
SHA-1 was offset introduced in 1995, and was already on its fashion out by the time Bursztein started working on this hash collision project with his colleagues in 2022. Past 2022, companies were beingness advised to apply SHA-two or SHA-iii instead. It would exist easy to dismiss the SHA-1 attack as a mere academic exercise, simply Bursztein was emphatic this wasn't the case. For instance, Mozilla, Microsoft, Google, and Apple all moved away from accepting SHA-1 ahead of schedule due to this research.
But the stakes are higher. Bursztein told the oversupply about how another team found a hash collision using a unlike cryptographic hash function called MD5 in 2009. Similar the SHA-1 standoff, it was an bookish success. Merely three years later, the Flame malware emerged in Iran. Notably, Flame used an MD5 hash collision to sign a artificial Windows update, meaning information technology tricked the figurer into thinking the malware was coming from Microsoft and could be trusted.
Bursztein conceded that this was the only known example of a weaponized hash collision, only there'due south a twist. The published methodology for an MD5 standoff used two blocks from the file to create a standoff. Forensic analysis revealed that the collision used in Flame used iv blocks. "So someone, somewhere, well funded, had developed their ain mode to create SSL collision," said Bursztein, who encouraged the audience to imagine what system would have that adequacy and be interested in infecting Iranian computers.
Despite this, Bursztein maintained that there is not cause for despair. There are numerous, more than secure hash functions at present available: SHA-256, SHA-iii, and BLAKE. Moreover, his team's research means that in that location are now tools available that can detect hash collisions, and that these tools are already in use in Github and Gmail, among other places. And with the rise of potential quantum computing attacks on cryptographic systems, robust cryptography and useful countermeasures are more than of import than ever.
Source: https://sea.pcmag.com/news/16754/researchers-reveal-secrets-of-sha-1-hash-collision
Posted by: milessuar1975.blogspot.com

0 Response to "Researchers Reveal Secrets of SHA-1 Hash Collision"
Post a Comment