Topic: Vitalik Buterin's idea of block size management is highly worth coinsidering (Read 448 times)
May 16, 2018, 05:44:38 PM
Vitalik is definately a prominent influencer, and keeping up with him can help people stay on top of new developments that drive the entire industry. It'd be worth it to check him out at this years TechCrunch Sessions on July 6th!
September 30, 2015, 08:03:52 PM
Ethereum Founder Vitalik Buterin Talks Cryptocurrency Governance
"To discourage a gas limit that leads to high mining centralization, Buterin proposed a target uncle rate. An “uncle” is kind of like an orphan block in Bitcoin—a valid block that doesn't make it into the main chain. So if the target uncle rate were 0.4, it would mean that the network is striving for 4 % of blocks uncled on average, explained Buterin. If the network were uncling at a higher rate, like 0.6, then miners would need to push down the gas limit. The gas limit has this effect because smaller blocks propagate faster and less uncling supposedly occurs in a network where propagation times are less varied."
"Buterin agreed, explaining that Ethereum benefits from development centralization in these early stages. But he doesn't expect that arrangement to last forever. “I think in the long-term something that's highly-decentralized is necessary,” he said. "
---------------------------------------------------------------------------------------------------------------------
Following his thoughts, for bitcoin, the top priority is decentralization, and a good indicator to measure decentralization is orphan rate. If the blocks process/propagate too slow, orphan rate will rise, indicating centralization risk
So, we can set a range for orphan rate, when orphan rate for the latest 2016 blocks is constantly lower than this range, it can be regarded as the block propagation is enough fast thus achieving full decentralization, then the block size can be raised by 10% following the difficulty re-adjustment
But there is a catch: If orphan rate is constantly low, block size will be raised to 20+MB but all the blocks will have little transactions in them. A 20MB block will create lots of problem under an attack, so there should also be another trigger condition, for example, average block being more than 75% full before raise the block size
In this way, if the PC/network infrastructure can handle, when average block is 75% full, the block size will be raised, so that blocks never become full under normal usage, the fee will be kept enough low while still ensure decentralization
However, if the PC/network infrastructure can not handle larger blocks due to higher orphan rate, then block size will not be raised because of the top priority is decentralization. Blocks will inevitably become full, and trigger the fee market mechanism so that other clearing based solutions will be more used to further decrease the fee for micro transactions. This will continue until the infrastructure of major miners improve
"To discourage a gas limit that leads to high mining centralization, Buterin proposed a target uncle rate. An “uncle” is kind of like an orphan block in Bitcoin—a valid block that doesn't make it into the main chain. So if the target uncle rate were 0.4, it would mean that the network is striving for 4 % of blocks uncled on average, explained Buterin. If the network were uncling at a higher rate, like 0.6, then miners would need to push down the gas limit. The gas limit has this effect because smaller blocks propagate faster and less uncling supposedly occurs in a network where propagation times are less varied."
"Buterin agreed, explaining that Ethereum benefits from development centralization in these early stages. But he doesn't expect that arrangement to last forever. “I think in the long-term something that's highly-decentralized is necessary,” he said. "
---------------------------------------------------------------------------------------------------------------------
Following his thoughts, for bitcoin, the top priority is decentralization, and a good indicator to measure decentralization is orphan rate. If the blocks process/propagate too slow, orphan rate will rise, indicating centralization risk
So, we can set a range for orphan rate, when orphan rate for the latest 2016 blocks is constantly lower than this range, it can be regarded as the block propagation is enough fast thus achieving full decentralization, then the block size can be raised by 10% following the difficulty re-adjustment
But there is a catch: If orphan rate is constantly low, block size will be raised to 20+MB but all the blocks will have little transactions in them. A 20MB block will create lots of problem under an attack, so there should also be another trigger condition, for example, average block being more than 75% full before raise the block size
In this way, if the PC/network infrastructure can handle, when average block is 75% full, the block size will be raised, so that blocks never become full under normal usage, the fee will be kept enough low while still ensure decentralization
However, if the PC/network infrastructure can not handle larger blocks due to higher orphan rate, then block size will not be raised because of the top priority is decentralization. Blocks will inevitably become full, and trigger the fee market mechanism so that other clearing based solutions will be more used to further decrease the fee for micro transactions. This will continue until the infrastructure of major miners improve
Jump to: