Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ethereum London Mainnet Announcement

blog.ethereum.org

Post reply on HN
Print view View original

71–80 of 153 posts

Re: Ethereum London Mainnet Announcement

#71

Post by wyxuan »

randomhodler84 wrote: post #13

Will the miners accept 1559? It’s an optional space, miners are free to maintain the old chain and split. Of course this would be catastrophic. I know my history, this isn’t anything like the other times.

Difficulty bomb ensures that forking is incredibly difficult.

The code change to defeat the code change is easy to make (just delete a section of code) but you have to distribute clients that ship with the difficulty bomb removed on top of EIP 1559 not being implemented, on top of people actually downloading the changed version.

Good luck trying to convince the Geth or other client teams to ship that version.

Re: Ethereum London Mainnet Announcement

#72

Post by nestlequ1k »

I understand some of the upgrades going into the London fork, including EIP-1559, but I don't understand "the merge". Does it mean that some kind of sharding will be in place, and the currently useless PoS chain will be able to communicate with the PoW chain?

Re: Ethereum London Mainnet Announcement

#73

Post by uncletammy »

wmf wrote: post #52
strombofulous wrote: post #36

Earlier quoted context omitted.

In think there's a subtle difference. A side effect of Bitcoin's rules is that people with more hashing power have a disproportionate amount of weight in deciding the future of the protocol. Eth solves this by putting a hard deadline on some replacement being chosen. This can be the current maintainer's updated fork, or it can be a fork someone made that's identical to eth today with the date pushed back a few more y…

95% of Bitcoin miners voted to increase the block size and it didn't happen, but you're right about the benefit of the ice age.

A thousand times this

Re: Ethereum London Mainnet Announcement

#74

Post by yokem55 »

jlokier wrote: post #65
3np wrote: post #41

Earlier quoted context omitted.

If you use spinning rust for the initial sync you’ll most likely never catch up. If you have craploads of RAM to allocate to ARC, you may actually be able to do it on ZFS I guess.

I think this may no longer be true with the Snap sync protocol (introduced in Geth this year). Even though transactions are very random access, Snap iterates sequentially over storage to transfer it all from one node to another, which is HDD friendly. However I'm not confident it will be able to stay in sync afterwards using HDD.

The new fork of geth, called erigon (formerly turbogeth) supposedly can do a full archive sync on a spinning disk drive and is much more space efficient.

Re: Ethereum London Mainnet Announcement

#75

Post by jl2718 »

dcolkitt wrote: post #31
liquidise wrote: post #2

EIP 1559 is a substantial change to the handling of gas fees on the network. In addition to making gas fees be 2 parts (base and "priority" fees), this includes a provision for "burning" the base fee, reducing the overall ETH supply. From the code itself: # miner only receives the priority fee; note that the base fee is not given to anyone (it is burned) Models i have seen suggest this will not make ETH deflationary,…

One really underrated component of EIP-1559 is how much more usable it makes the system. The complexity and uncertainty of how to set gas fees, stuck transactions, and the like is a major turnoff for new users. Experienced users tend to gloss over this, because they've already taken the time to mentally model all the complexities related to the gas market. I predict that post-London, we'll see a huge wave of new user…

All of that complexity is still in the priority fee. There is still zero incentive to include base-fee-only transactions.

Re: Ethereum London Mainnet Announcement

#76

Post by everfree »

nestlequ1k wrote: post #72

I understand some of the upgrades going into the London fork, including EIP-1559, but I don't understand "the merge". Does it mean that some kind of sharding will be in place, and the currently useless PoS chain will be able to communicate with the PoW chain?

Basically, at the point of the merge, nodes stop paying attention to what the PoW calculations are saying about the head of the execution chain, and start to pay attention to what the beacon chain (PoS chain) is saying about the head of the execution chain.

Because the two chains become inextricably linked together at that point, with the execution chain becoming entirely dependent on the beacon chain, we call that point the "merge".

So from that point on, new blocks on the execution chain - full of general Ethereum transactions - will be considered "confirmed" not because they include a difficult PoW calculation, but because the beacon chain has given that block its blessing (because stakers voted on it).

Re: Ethereum London Mainnet Announcement

#77

Post by lupire »

dcolkitt wrote: post #39
dsr_ wrote: post #33

Earlier quoted context omitted.

That implies that USDCblah is actually backed by traceable assets, 1:1. Have any of the so-called stable coins actually passed a complete third-party asset?

That's largely an orthogonal point. You may be suspicious of USDT or USDC's claim. But the point is that at least some fraction of the market is not, which is why USDT trades at $1.00. Now that may turn out in the future to be revealed to be a sham, and the market will lose faith. But right now USDT only has market value because some segment of people believe in the promises made by the Tether Corporation. If I fork…

not quite. stablecoins do not get stability from belief in the backing assets. they get stability merely from a shared agreement to pretend the coin is stable, so as not to lose whatever stake they have put in.

Re: Ethereum London Mainnet Announcement

#78

Post by lupire »

cinquemb wrote: post #56
wmf wrote: post #53

Earlier quoted context omitted.

That would explicitly break the one promise they made, which is that their coins are backed 1:1.

Frax, USDT, and TrueUSD allow people to deposit into bank accounts and mint on eth mainnet or avalanche c-chain now (note: not mint on both places at the same time), neither make the claim that they are 1:1 backed (except for TrueUSD), but they make the claim that if you wanted to at any time (if you meet the criteria) you can withdraw from eth mainnet or avalanche c-chain to your bank account.

When did Tether USDT stop (fraudulently) claiming to have backing assets?

Re: Ethereum London Mainnet Announcement

#79

Post by lupire »

N00bN00b wrote: post #67
doomroot wrote: post #21

Earlier quoted context omitted.

Doesn’t that seriously centralize the network overtime?

We'll see. Stablecoins are a workaround for the current price fluctuations, if the cryptos stabilize they might replace the stablecoins at some point in the future.

If cryptos stabilize 99% of users will lose interest.

Re: Ethereum London Mainnet Announcement

#80

Post by lupire »

jlokier wrote: post #66
tomc1985 wrote: post #32

Earlier quoted context omitted.

I don't understand, why is spinning metal not good enough? For the large amount of storage required it seems like it would be a good idea.

It's a great idea. It just doesn't work. HDDs don't have the right performance characteristics, unfortunately. You can do about 300 random access per second on a good HDD, which isn't enough to keep up with Ethereum's transaction rate, and each transaction needs quite a few of those each. You can do about 300,000 per second on NVMe. 1000 times faster. That's plenty.

what happens when eth get popular and nvme can't keep up?
Post reply on HN