[Cryptography] DAG vs Blockchain

Ersin Taskin hersintaskin at gmail.com
Sun Jan 28 16:37:16 EST 2018


 On 01/26/2018 06:09 PM, bexnews at gmail.com wrote:
> Much hype exists around DAG (directed acyclic graph) in some parts, as
> an improvement on blockchain crypto because, for one thing, it does away
> with mining. However isn't mining the thing that incentivizes increased
> computing power to verify the blockchain and protect against a 51%
> attack?  Would DAGs be less secure against attackers with lots of
> computing power *because* there is no mining?

2018-01-27 19:45 GMT+03:00 Valentin Pasquale <valentin.pasquale at orange.fr>:

> "However isn't mining the thing that incentivizes increased computing
> power to verify the blockchain and protect against a 51% attack?"
>
> Actually, you can "mine" with stack of coin (that's PoS of dPoS). Ethereum
> is trying to implement this kind of mining
>
> "Would DAGs be less secure against attackers with lots of computing power
> *because* there is no mining?"
>
> Same, if a system is based on something (let say mining) then yes dag is
> easier to attack. With hashgraph you need to be 34%. But some people claim
> every distributed system is vulnerable to a 34% attack, consisting in
> setting up a firewall separating honest miners in two groups, and then
> attack of the group. Such a firewall can be pretty hard to set, so
> blockchain is more secure.
> But dag is designed to be "fast" and therefore isn't based on mining but
> rather on stack or reputation (for permissioned systems, a group of
> founders can have 20% of the reputation they share when they invite people
> in). Moreover I think proof of stake or proof of work achieve the same
> level of security regardless of how one can attack it (PoS can be insecure
> if they are massive hacks, and PoW can be insecure if one guy achieve to
> improve the "mining algorithm" and sell it. Personally I prefer proof of
> work)
>
> TLWR : "No". Yes.
>


Pow, pos, dpos, etc. all have the same problem when it comes to security.
When you are in the ocean, you cannot anchor to something made of or
convertible to water. If you anchor to an iceberg, the attackers just melt
it. In a financial system, you cannot anchor to anything made of or
convertible to money. Pos, pow, etc suffer because stake is made of
money, electricity is convertible to money. So all the cryptocurrencies are
vulnerable to a super-rational attack involving enough stake/power. I call
this the AR Theorem (Anchorlessness-Robustness) in my paper. I show that
you cannot achieve both Anchorlessness and Robustness at the same time. I
define Robustness with components like efficiency, super-rational attack
resilience, etc. Funnily, we have an anchorage that we amazingly ignore. An
anchorage you cannot buy, bribe, corrupt. An anchorage that everybody on
this list trust. An anchorage that can be used to randomly create and use
secret keys that secures the system. It is right there next to us. We just
don't see it because we don't raise our heads. And we keep paying an
amazingly huge technical debt to get vulnerable and poor systems:) This is
a manifestation of the importance of skepticism, the fuel of science. Zero
tolerance to fanaticism. We, the scientists, should never give up being
skeptical about the things that we believe the most. That is the essence of
paradigm shifts, which brought humanity here, whose frequency needs to
increase.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20180129/06b50801/attachment.html>


More information about the cryptography mailing list