Constantinople Now Scheduled for 27th February, ProgPoW Decision Seemingly Reversed

0

Ethereum developers have now set a new block number for Constantinople which is to occur at block 7,280,000, estimated to be on or around 27th of February.

That’s in about six weeks from now due to further time needed for testing as the SSTORE net gas metering Ethereum Improvement Proposal (EIP) is to be removed.

There was some discussion on whether SSTORE can be salvaged for this fork, but with another one planned for October, developers decided it is just better to include it in Istanbul.

Constantinople, therefore, is to go ahead with a block reduction to two eth, a difficulty bomb delay for 12 months, and one or two other somewhat minor protocol improvements.

Developers had further discussions regarding a proposed algorithmic change from ethhash to ProgPoW with their conclusion being that the dev call was  not the right forum to make a decision.

There were two major objections to ProgPoW, although expressed very politely and framed as not explicitly being an objection.

Alexey Akhunov, who is working on rent storage and other improvement proposals for Ethereum 1.x, said he was not aligned with asics or no-asics, but from his point of view concerning his task of developing and designing eth 1.x, the more hardforks there are planned, the harder for eth1x.

Discussion followed in regards to the Proof of Stake (PoS) Beacon Chain with Vitalik Buterin saying it was unrealistic to expect it within six months after stating that the Beacon chain can be usable as a finality gadget – as original Hybrid Casper.

In that case, he said, and if enough people participate in PoS to the point you can say the finality gadget aspect is secure, then a 51% attack will only be able to censor, but not reverse blocks.

Upon an anonymous individual who has been pushing for ProgPoW, Mr. Def, asking whether in such a situation block censoring wouldn’t be a problem, Buterin effectively said we’d just brick them.

If there is censorship, then the only thing we can do if we know they are using asics hardware is to change the Proof of Work (PoW) algorithm, Buterin said. While if it is gpu hardware, he added, we would have no choice but to scramble to migrate to PoS.

“If attacker using gpu, there’s nothing,” Buterin said, with discussion following whether fork-choice rules can be modified in such situation. Something which, Buterin said, may split the network in half and wreck huge havoc unless it is somehow tied to the 99% fault tolerant method.

The anonymous individual, Mr Def, finally “officially” admitted in the call that Nvidia gear is favored in two ways.

There’s apparently some AMD compiler bug, he claimed. They are engaging with AMD engineers to see how to mitigate the issue. Moreover, the computation requirements for AMD are a “little bit too harsh , too compute heavy.” They are now trying to “turn it down a bit, helps some of the AMD hardware.”

Finally, ethereum has cat herders now. According to Hudson Jameson and Lane Rettig, the cat herders are a group of 12 people who take notes, take surveys. They are volunteers with real project management experience tasked with, basically, project management.

There was also some talk about how a fork can be pulled off at the last minute. Péter Szilágyi didn’t suggest, but merely put out there this idea of a centralized Oracle, kind of basically a kill off switch.

The considered scenario was what if some bug was found say 3 hours before a fork, rather than 36. Nakamoto’s solution to that was an alert key. Basically a word/typed message showing at the corner of the node saying this has happened, do this, more info wherever.

The necessarily English centric aspect of such message, as you probably can’t write more than so many words on it, may cause problems but with Google Translate now getting pretty good at its job, the language of the linked article might not quite matter.

They’ve removed the alert key in bitcoin however because they consider it a centralizing aspect in some way as there would be a select group of people communicating with all the nodes.

It’s unclear, however, whether that would be a worthy trade-off until sharding and so on, until a point where we can say this is now settled in stone.

Yet it is probably far better than any oracle as it would inform the nodes while giving them the free choice of action. Including, if they want to, their own go ahead with the called off fork.

That’s what some miners did yesterday in regards to Constantinople. That appears to have been just accidental, with things now back to normal under one eth chain that is currently undergoing the ice age.

Copyrights Trustnodes.com

 

Comment

100000
  Subscribe  
Notify of