Besides I have explained how DPoS is vulnerable to transaction spam attack per Vitaliks point
nonsense. [
]
Here's the blue part you highlighted, and it's going off wrong assumptions. he assumes minimum transaction bandwidth is bad.
The bolded point above by Vitalik is that if transactions are free up to the burst limit, then the DoS attackers can avail of it. Thus the DPoS blockchain is attacked and not just the servers (nodes) on the periphery. A sock puppet attack can be employed to defeat any attempt by the block producers to fairly limit each user. This would have the effect of forcing all users to the minimum transaction bandwidth their stake will accord, because a significant portion of the stake is maxing out the bandwidth and compensation of the block producers.
The burst capacity is of limited use to spammers. There is a burst limit that is some moderate multiple of your minimum allocation, I think currently 10x. If you have 0.1% of the stake then you can use up to 1% of the bandwidth, burst capacity permitting. Even if you spam at the max rate, that still leaves 99% available for others to share. If the blocks start to fill up then the overcommit ratio drops. A 0.1% stakeholder would then only be able to fill up hypothetically 0.5% if the overcommit ratio were cut in half. (I don't know the exact algorithm governing it.)
You would need to own a meaningfully large portion of the stake (10% or more) to force users to burst at their minimum allowance, and even then the overcommit ratio would dynamically drop and then you would only effectively limit them to a smaller multiple of their minimum allowance.
I do not understand why both of you guys seem to be ignoring Vitaliks point.
As I understood his point to be that as spammers max out the utilization (presuming they have significant stake, not the 0.1% in your contrived example) then users no longer have that reserve of burst capacity available, thus they have to buy exactly what they may need in their 99th percentile usage pattern and this is thus more expensive than just paying transaction fees.
Granted that Vitaliks scenario does require the attacker to control significant stake (and spread it around in sock puppets), and you might retort that why would those who own a majority of the state want to destroy the value of their own coins. Yet you @smooth afair used to argue exactly the opposite against proof-of-stake arguing that since nothing is at stake they can short the market, profit, rinse, buy low, the market rebounds on their pump, then repeat. Unless I am somehow misstating your past stance on proof-of-stake.
Sorry the inconsistency around here is wasting my time.
I wasnt paid to explain Vitaliks thinking. I get paid to produce an altcoin. And so please excuse me if I exit from this never ending series of explaining everything.
I have no beef with @smooth. I appreciate very much our interactions
(weve had our instances of strife which can be due to numerous factors but overall is dominated by rational dialogue). I have a beef with those who use @smooth as some battering ram against me.
A sock puppet attack would do nothing since bandwidth is divided from maximum bandwidth per vested coin, not per account, so irrelevant.
The point is that if any attempt was added to try to filter by account, it could then be subverted with a sock puppet attack.