@multifolio @durerus @kabizses 1\ IOTA will not use the same FIFO mechanism as a tie breaker in times of congestion as ETH, where we just use mana instead of fees to decide which transactions to process.
Instead, we will use an agent-centric sharding approach where we use the "distance" of transactions to ...
Instead, we will use an agent-centric sharding approach where we use the "distance" of transactions to ...
@multifolio @durerus @kabizses 2\ ... the corresponding nodes to dynamically react to differences in supply and demand.
If nodes can not process enough transactions anymore, then they will reduce their observation radius until they are able to keep up again.
See medium.com/@hans_94488/sc… for a description.
If nodes can not process enough transactions anymore, then they will reduce their observation radius until they are able to keep up again.
See medium.com/@hans_94488/sc… for a description.
@multifolio @durerus @kabizses 3\ This mechanism alone is however not enough as this would allow an attacker to just randomly spam a certain shard, forcing the nodes to shrink their observation radius and reducing the security in the network.
We therefore still have a mana-based rate control mechanism on ...
We therefore still have a mana-based rate control mechanism on ...
@multifolio @durerus @kabizses 4\ ... top of that where we observe the activity of the n-percentile of the mana-holding nodes in the network and are only shrinking the perception radius if the majority of honest nodes are really requesting this additional throughput.
In a "saturated" network, this still ...
In a "saturated" network, this still ...
@multifolio @durerus @kabizses 5\ ... establishes sth like a "minimum mana-requirement" for nodes to issue transactions preventing things like attacks or people using colored coins instead of the native tokens.
This minimum requirement is however no fixed threshold but will naturally evolve as an ...
This minimum requirement is however no fixed threshold but will naturally evolve as an ...
@multifolio @durerus @kabizses 6\ ... equilibrium of the economic activity in the region and the distribution of reputation among nodes.
This minimum-threshold will ideally be pretty low so no honest actor will ever be blocked from using the network, but high enough to prevent these kind of "attacks".
This minimum-threshold will ideally be pretty low so no honest actor will ever be blocked from using the network, but high enough to prevent these kind of "attacks".
@multifolio @durerus @kabizses 7\ The exact parameters and formulas around this are still part of research as they also depend on things like "real world mana distribution".
We suspect that the mana and "wealth" distribution follows "Zipf's law" and we therefore have quite good "guestimates" but this is ...
We suspect that the mana and "wealth" distribution follows "Zipf's law" and we therefore have quite good "guestimates" but this is ...
@multifolio @durerus @kabizses 8\ ... something that is best observed by studying the mainnet after coordicide and then adjusting the parameters accordingly as soon as we introduce sharding.
Gladly there have been decades of research into these kind of rate-control mechanisms in networks like AIMD, so we ...
Gladly there have been decades of research into these kind of rate-control mechanisms in networks like AIMD, so we ...
@multifolio @durerus @kabizses 9\ ... don't have to reinvent the wheel :P
I hope this makes it a bit more clear - if not just PM me in discord and we can go a bit more into the details.
I hope this makes it a bit more clear - if not just PM me in discord and we can go a bit more into the details.
