Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

BitTorrent v2 (2020)

blog.libtorrent.org

Post reply on HN
Print view View original

21–30 of 155 posts

Re: BitTorrent v2 (2020)

#21

Post by marcan_42 »

elcomet wrote: post #16
hannob wrote: post #15

Earlier quoted context omitted.

You're proposing to add extra complexity for some hypothetical scenario that is unlikely to happen. That's a bad tradeoff.

unlikely ? all methods are broken eventualy when compute becomes faster

This is not true. Modern cryptography is not breakable via brute force any more, nor will it ever be. You can prove that the amount of energy required assuming a thermodynamically ideal computer is more than will ever be available in our galaxy, with classical computing.

Breaks of symmetric crypto or hashes that are practical require an actual cryptographic weakness, whose existence is not a given.

Then there's quantum computing, but that also doesn't break symmetric crypto outright, it just makes it weaker. I'm not sure if anyone has run the math, but we'd probably still be talking galaxy sized quantum computers to get anywhere.

Re: BitTorrent v2 (2020)

#22

Post by alophawen »

hannob wrote: post #15
dennis-tra wrote: post #12

I‘m wondering if they could avoid yet another breaking protocol change if SHA256 proves to be insecure (at some point in the future) if they made use of Multiformats [0]. At least IPFS went this way (Multiformats grew out of the IPFS development). [0] https://multiformats.io/

You're proposing to add extra complexity for some hypothetical scenario that is unlikely to happen. That's a bad tradeoff.

There's nothing unlikely with such scenario. See MD5, SHA1 and so on.

See current research on SHA256

https://en.wikipedia.org/wiki/SHA-2#Cryptanalysis_and_valida...

https://dusted.codes/sha-256-is-not-a-secure-password-hashin...

https://security.stackexchange.com/questions/34256/sha256-se...

Re: BitTorrent v2 (2020)

#23

Post by elcomet »

marcan_42 wrote: post #21
elcomet wrote: post #16

Earlier quoted context omitted.

unlikely ? all methods are broken eventualy when compute becomes faster

This is not true. Modern cryptography is not breakable via brute force any more, nor will it ever be. You can prove that the amount of energy required assuming a thermodynamically ideal computer is more than will ever be available in our galaxy, with classical computing. Breaks of symmetric crypto or hashes that are practical require an actual cryptographic weakness, whose existence is not a given. Then there's quant…

thanks, that's interesting, I didn't know this !

Re: BitTorrent v2 (2020)

#24

Post by seanw444 »

hannob wrote: post #20
elcomet wrote: post #16

Earlier quoted context omitted.

unlikely ? all methods are broken eventualy when compute becomes faster

Computers can only become faster within the limits of what humans are able to build and what physics allows. You're not gonna break SHA256 with faster computers.

I swear there was a pertinent article link that goes through how it's physically impossible to construct a computing device fast enough to break SHA-256, even at the most fundamental scale.

Re: BitTorrent v2 (2020)

#25

Post by hannob »

alophawen wrote: post #22
hannob wrote: post #15

Earlier quoted context omitted.

You're proposing to add extra complexity for some hypothetical scenario that is unlikely to happen. That's a bad tradeoff.

There's nothing unlikely with such scenario. See MD5, SHA1 and so on. See current research on SHA256 https://en.wikipedia.org/wiki/SHA-2#Cryptanalysis_and_valida... https://dusted.codes/sha-256-is-not-a-secure-password-hashin... https://security.stackexchange.com/questions/34256/sha256-se...

The Wikipedia link is very reassuring regarding the security of SHA256. All attacks are very far away from going anywhere close to attacking full SHA256 and no major progress has been made lately.

The password hashing remark is correct, but irrelevant. SHA256 is not a good password hash, because it's not made to be a password hash. Don't use it for passwords.

The weaknesses that broke MD5/SHA1 were known since 1994. No similar weakness is known for SHA256.

Re: BitTorrent v2 (2020)

#26

Post by samstave »

marcan_42 wrote: post #21
elcomet wrote: post #16

Earlier quoted context omitted.

unlikely ? all methods are broken eventualy when compute becomes faster

This is not true. Modern cryptography is not breakable via brute force any more, nor will it ever be. You can prove that the amount of energy required assuming a thermodynamically ideal computer is more than will ever be available in our galaxy, with classical computing. Breaks of symmetric crypto or hashes that are practical require an actual cryptographic weakness, whose existence is not a given. Then there's quant…

This is why they create laws to require you unlock your phone when crossing borders.

I have been constantly surprised that there isnt a phone/company that specifically markets to "buy this temp phone to cross a border"

--

Also, WRT Quantum computing, what are your thoughts on the talk by D-Wave CEO where he said "we can reach into different dimensions" -- and soon after, D-Wave went super silent?

What is the state of quantum currently - did they discover something super secret?

Re: BitTorrent v2 (2020)

#27

Post by alophawen »

hannob wrote: post #25
alophawen wrote: post #22

Earlier quoted context omitted.

There's nothing unlikely with such scenario. See MD5, SHA1 and so on. See current research on SHA256 https://en.wikipedia.org/wiki/SHA-2#Cryptanalysis_and_valida... https://dusted.codes/sha-256-is-not-a-secure-password-hashin... https://security.stackexchange.com/questions/34256/sha256-se...

The Wikipedia link is very reassuring regarding the security of SHA256. All attacks are very far away from going anywhere close to attacking full SHA256 and no major progress has been made lately. The password hashing remark is correct, but irrelevant. SHA256 is not a good password hash, because it's not made to be a password hash. Don't use it for passwords. The weaknesses that broke MD5/SHA1 were known since 1994.…

If you want to be reassured then sure. My take from reading that is in 2016 there was practical collision attacks demonstrated, which would be amplified with specialized hardware (compare to bitcoin mining).

Btw cheers! I recall you from correspondence years ago regarding fuzz testing linux packages.

Re: BitTorrent v2 (2020)

#28

Post by hffftz »

pndy wrote: post #2

https://news.ycombinator.com/item?id=24401999 - 2020, 1227 points, 562 comments

It would help if HN would automatically link to previous discussions (next to the title?) if there are any... It seems wasteful having people do it manually

Re: BitTorrent v2 (2020)

#29

Post by xoa »

marcan_42 wrote: post #21
elcomet wrote: post #16

Earlier quoted context omitted.

unlikely ? all methods are broken eventualy when compute becomes faster

This is not true. Modern cryptography is not breakable via brute force any more, nor will it ever be. You can prove that the amount of energy required assuming a thermodynamically ideal computer is more than will ever be available in our galaxy, with classical computing. Breaks of symmetric crypto or hashes that are practical require an actual cryptographic weakness, whose existence is not a given. Then there's quant…

>Modern cryptography is not breakable via brute force any more

Specifically, modern symmetric crypto (and hashing which is a related but different thing) is fine. It's public-key crypto that has always been the worry thanks to Shor's Algorithm, and all modern widely used crypto there is indeed vulnerable if an actual scalable general quantum computer could be constructed. There are a variety of post-quantum cryptographic algorithms under development including some ones that would be ugly slow but likely effective bandaids if required, but that's still a not totally unreasonable concern for certain threat profiles. But yes GP was totally wrong about "all methods are broken when compute becomes faster".

>I'm not sure if anyone has run the math

"The math" here is just Grover's Algorithm, that lowers the cost of generic brute forcing to O(sqrt(n)). So a 128-bit symmetric key could be cracked on average like 2^64 or a 256-bit key like 2^128. Obviously this is utterly trivially countered by doubling the exponent. Even 2^128 is still ridiculous, and going to a 512-bit key brings us right back to 2^256 which is impossible. 128-bit keys should indeed be phased out entirely (and that seems to be well on the way anyway), and perhaps 256-bit too at some point (even if it's only the very paranoid or very long termers, it's also not big stretch on modern hardware at all, so eh), but fundamentally yes symmetric crypto is fine.

Re: BitTorrent v2 (2020)

#30

Post by antris »

samstave wrote: post #26
marcan_42 wrote: post #21

Earlier quoted context omitted.

This is not true. Modern cryptography is not breakable via brute force any more, nor will it ever be. You can prove that the amount of energy required assuming a thermodynamically ideal computer is more than will ever be available in our galaxy, with classical computing. Breaks of symmetric crypto or hashes that are practical require an actual cryptographic weakness, whose existence is not a given. Then there's quant…

This is why they create laws to require you unlock your phone when crossing borders. I have been constantly surprised that there isnt a phone/company that specifically markets to "buy this temp phone to cross a border" -- Also, WRT Quantum computing, what are your thoughts on the talk by D-Wave CEO where he said "we can reach into different dimensions" -- and soon after, D-Wave went super silent? What is the state of…

They? Who, where? Which law?
Post reply on HN