Vitalik: Sharding is coming

[satrum]

it is time price follow up?
News added for Verification on YEENOT TODAY  https://yeenot.today/catalog/news/162

github:
https://github.com/ethereum/research/tree/master/sharding_fork_choice_poc

This is a proof of concept of (part of) a fork choice rule-based mechanism for how sharding can be bolted on top of the current ethereum main chain, with a specialized random beacon and shard block times of <10 seconds. The basic idea is based on a concept of dependent fork choice rules. First, there is a proof of stake beacon chain (in phase 4, aka full casper, this will just be merged into the main chain), which is tied to the main chain; every beacon chain block must specify a recent main chain block, and that beacon chain block being part of the canonical chain is conditional on the referenced main chain block being part of the canonical main chain.

The beacon chain issues new blocks every ~2-8 seconds, with a design similar to the one prototyped here (implementation at https://github.com/ethereum/research/tree/master/old_casper_poc3), using the RANDAO mechanism to generate randomness (see https://ethresear.ch/t/rng-exploitability-analysis-assuming-pure-randao-based-main-chain/1825, https://ethresear.ch/t/rng-exploitability-analysis-assuming-pure-randao-based-main-chain/1825/10 and http://vitalik.ca/files/randomness.html for analysis of exploitability for analysis), and its purpose is to be the "heartbeat" for the shard chains and to provide the randomness that determines who the proposers and notaries in the shard chains are. The beacon mechanism is upgraded with a proof of activity-inspired technique to increase its stability.

The shards then themselves have a dependent fork choice rule mechanism that ties into the beacon chain; every time a new beacon block is created, that beacon block randomly selects a proposer which has the right to create a shard collation. Each shard collation points to a parent collation on the same shard, and a beacon block.

Things that are not included in this test are:

    The mechanism for notaries to confirm shard collations (though this is trivial to implement; it's the same as for beacon blocks)
    The shard-to-main-chain crosslink (see https://ethresear.ch/t/cross-links-between-main-chain-and-shards/1860) that ties the beacon and the shard chains back into the main chain
    The feature where all notarizations of any shard simultaneously double as votes in a global Casper FFG cycle, increasing Casper FFG scalability and allowing its min deposits and finality times to both be reduced (perhaps min deposits to 32 ETH and finality times to ~6 minutes)

[BitcoinNewbie15]

Its great to see some progress in the sharding tech! A question I have is does anyone know how significant the improvement would be for ethereums throughput once sharding is put into effect? In terms of TPS? Also how exactly would Sharding effect decentralization? I was under the impression it could potentially increase centralization, but I am not too keen on the sharding technology.

[Judster]

And you see negative effects on equity valuations of companies that develop similar projects. And down they went, you know.
I  didn't fully understand the intricacies. I do not have enough technical knowledge. But Vitalik is a fine fellow anyway. Generating ideas and translating them into reality is always very cool.

[FrueGreads]

I've heard of this before, and from a basic perspective I believe this will do for ethereum what the lightning network is doing for bitcoin, solving it's scalability issues, and making transactions a hole lot faster. But from what I've last heard, and it seems that this is mentioned here again, this will only be done with that casper update, that is suppose to be the "final step" on ethereum roadmap (it will probably have more updates of course, but from what I know, this is the last stage of already planned development phases).

I'm just mentioning this, because I believe we still need to wait a few years for casper, so I don't expect this to drive the price up right now.

[boy130]

Looks like Ethereum is adopting a Zilliqa light strategy by implementing sharding. I guess that's one of the major problems with new cryptocurrencies when they are based on the ERC-20 standard. Ethereum can simply adopt all of the best features of any cryptocurrency using its ERC-20 standard, hence they can never overtake ETH as the number two coin. Similar to Bitcoin which can just keep forking to keep up with the new generation coins, but it does appear to be quite resistant to forking unlike ETH which undergoes regular protocol upgrades.

[FrueGreads]

Looks like Ethereum is adopting a Zilliqa light strategy by implementing sharding. I guess that's one of the major problems with new cryptocurrencies when they are based on the ERC-20 standard. Ethereum can simply adopt all of the best features of any cryptocurrency using its ERC-20 standard, hence they can never overtake ETH as the number two coin. Similar to Bitcoin which can just keep forking to keep up with the new generation coins, but it does appear to be quite resistant to forking unlike ETH which undergoes regular protocol upgrades.

I understand what you are saying and I do agree with part of it, but I'm not sure it's exactly like that for everything. For instance when bitcoin forked and bcash was born, they really took separate ways, so I don't think bitcoin would ever go back to what BCH was doing. Of course that bcash just increased the block size and bitcoin could still do that if they wanted to, but it wouldn't make any sense for them to do it.

What I'm trying to say is that in theory what you are saying might be correct, but "in the field" the community will probably get divided on important forks, and will not adopt each others strategies in the future.

[asriloni]

Looks like Ethereum is adopting a Zilliqa light strategy by implementing sharding. I guess that's one of the major problems with new cryptocurrencies when they are based on the ERC-20 standard. Ethereum can simply adopt all of the best features of any cryptocurrency using its ERC-20 standard, hence they can never overtake ETH as the number two coin. Similar to Bitcoin which can just keep forking to keep up with the new generation coins, but it does appear to be quite resistant to forking unlike ETH which undergoes regular protocol upgrades.

Bitcoin fork is really different compared with ethereum fork. In ethereum fork to generate or make another separate blockchain will be minimized and it's not like bitcoin to generate fork and fork anytime. Remember it has raiden network in another side and not only sharding. EIP1011 is on the way and this gonna be the biggest thing in the future.

[pinkflower]

This has been on their roadmap for more than 2 years if I remember it right. While its good to see some forward development, I think it will take 2 or more years to have it impelemented.

Will sharding will require another hard fork?