Apologies problem with eye sight earlier (no joke). Now I see the bounty rules link. So roughly $2100 at current exchange rate. Probably not worth my time right now, because I am working on something orders-of-magnitude more lucrative and under a deadline.
I changed the post in the other thread, so as to link to this thread and check for discussion here.
I am not referring to a 51% attack.
It appears there is some sort of "zerotime" algorithm wherein the P2P network attempts to form a consensus (a "global lock") on inputs to a 0-confirmation transaction sent to the network. I didn't study the details of the network propagation exhaustively. I see some diagram with "slots" and various nodes propagate a lock logic across the P2P network.
Well recently Skycoin has proven the game theory about Sybil attacks with reputation, propagation networks and the best case is that if 7% of the nodes lie, then the network fails to get the correct consensus:
I will quote from your white paper:
Under the SkyHash model the wiki dataset can survive under
DoS attack committed by 7% random nodes or 0.9% top
influential nodes defined as the first 0.9% nodes by sorting all
nodes in descendant order on the count of a node’s followees,
however, the throughput will decrease 50% even when the
network survives. In all the cases that the network survives,
correct nodes can always reach almost-everywhere consensus
within 45 seconds without correct nodes agree at different
values, while under DoS attack by 7% random nodes, 1.5%
nodes refuse to agree at any values, and under DoS attack by
0.9% top influential nodes, 4% nodes refuse to agree at any
values.
As we introduced in Section II, Bitcoin’s PoW is the
best Sybil-proof consensus at present, but it is a different
mechanism to our work and not comparable directly in Fig. 6.
Through the automatic adjustment of the difficulty of PoW,
Bitcoin generates a block in about 10 minutes, and a fully
confirmed consensus need 6 blocks thus needs about 1 hour.
However if a single node or a group of nodes has a large
proportion of compute power, it can compromise the network
and create a fork. Table II shows the probability of success
attack for 6 blocks confirmations [23]. If one adversary in
Bitcoin has a threatening compute power, the whole network
can’t do anything to resist it because the power is controlled
by the adversary itself, while in our approach a node’s power
is controlled by its followees, thus a node can be unarmed by
unfollowing it.
...
Rather the key innovations of your consensus algorithm appear to be:
- It is based on reputation following instead of compute power.
- The fail case is only 7% (or 0.9% for influential nodes) versus 50% (or actually as low as 25% for selfish-mining) for proof-of-work.
- It doesn't waste computing resources and electricity.
...
Hopefully you can understand the "global lock" is not leveraging any compute power for Byzantine fault resistance, i.e. it is not proof-of-work. Rather it is relying on the structure of the propagation of the P2P network to converge on a consensus. This is basically analogous to Skycoin's consensus network at the conceptual level. So the analysis of vulnerability should be in the same ballpark.
There are no short-cuts around proof-of-work or proof-of-share for unambiguous, global consensus, i.e. Byzantine fault resistance. If the developer of VanillaCoin thinks otherwise, he needs to prove it with some math. I don't see any such proofs in the white paper.
It is almost as if he didn't learn Bitcoin 101, which is ludicrous to say, because obviously he is knowledgeable. So either I have a big blind spot or he just somehow deluded himself. It is actually bizarre.
The only thing you could do is poll the nodes and make sure 51% of the network hashrate is reporting the global lock. If you can tie the reputation to hashrate, then you might have a solution. But then the problem is they can lie and you can't prove they won't (if necessary via a Sybil identity).
You've got three ways to do consensus: proof-of-work, proof-of-stake/share, or long-term reputation.
If it was as simple as he apparently thinks it is, don't you think it would have already been done for Bitcoin?
