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

Ethereum Plans to Cut Its Energy Consumption by 99 Percent

spectrum.ieee.org

Post reply on HN
Print view View original

271–280 of 347 posts

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#271

Post by jillesvangurp »

wslh wrote: post #225
jillesvangurp wrote: post #220

I'm glad we switched away from Ethereum in favor of Stellar; it will take most of this year and probably next for the dust to settle on this. Until then, Ethereum continues to be unusable for serious dapps (which are rare to begin with) that have expectations that exceed having reasonable transaction throughput ( Until then, this is a toy platform for wannabe ICOs, misc gambling applications, and not much else that I…

The Stellar protocol has not been yet peer reviewed. Not saying that is wrong but I would say that when you start relaxing consensus assumptions you can come up with even simpler protocols. For example, using trusted third parties to timestamp your transactions. In this way you solve the ordering issue in consensus (assuming no timestamp can give the same time). For example, imagine if Google, Amazon, Microsoft, and…

I understand; this has also been a concern for us. However, there's actually very little out there that is actually mature and usable. Bitcoin is not usable for custom applications period. It's nice if you want to trade bitcoin; it doesn't do anything else. Ethereum has enough of a user base that you could argue it is probably OK in terms of security; however as argued it's not really usable in its current form. Everything else is a combination of too young, technically challenged, too small in user base, etc.

Bearing this in mind, we picked Stellar based on the fact that it has actually been around for a while and launched late 2017, is used by several more serious players in the ecosystem (e.g. IBM seems to like it) and a few fin tech companies are building stuff with it. It has a few tokens that seem less about speculative trading and more about solving a real problem (e.g. several stable coins that tend to be relatively stable in value, by design).

If you look at e.g. EOS, the picture is a bit different. It seems to be favored by asian gambling tokens which have pathetic numbers of transactions per day. Also, they launched after we decided to move to Stellar. Tron seems to have similar issues. Then there are a lot of smaller chains that are currently not launched yet e.g. IOTA, hashgraph, that you might see as potentially of interest in the future. After that comes a long tail of solutions that nobody can vouch for right now.

So, we're taking some risks here with Stellar but it seems to be relatively OK for us so far and I haven't really seen anything more suitable for our needs so far. In terms of transactions per second there's way more happening on Stellar than Ethereum is even capable of handling currently. Also, Stellar seem to have distanced themselves a bit from the whole ICO business, which I count as a good thing.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#272

Post by DennisP »

keymone wrote: post #265
DennisP wrote: post #263

Earlier quoted context omitted.

That's true. There are several specific slashing conditions, and I forget the exact details, but they're all similarly well-defined and limited. There are proofs that the network converges properly as long as most stake obeys those conditions, and that's why only violations of those conditions are penalized.

> There are proofs that the network converges properly as long as people don't violate those conditions do you have a link to those proofs? i've seen proofs already where underlying assumptions were something like "majority of nodes are honest" or "network is reliable", etc.

I'm not up to date but here's the Casper FFG paper, which says:

> Accountable safety means that two conflicting checkpoints cannot both be finalized unless ≥1/3 of validators violate a slashing condition (meaning at least one third of the total deposit is lost). Plausible liveness means that, regardless of any previous events (e.g., slashing events, delayed blocks, censorship attacks, etc.), if ≥2/3 of validators follow the protocol, then it’s always possible to finalize a new checkpoint without any validator violating a slashing condition.

https://arxiv.org/abs/1710.09437

So if more than a third of the stake violates a slashing condition, that does cause a problem, but only a temporary one since that stake is immediately destroyed. (An important improvement over just assuming majority honest nodes is that you can prove who violated any of the slashing conditions.)

There's been a lot of work since that paper, much of it discussed at ethresear.ch.

Casper FFG is the version currently heading for production; in the paper it's described as a finality layer adding security to proof of work. The original plan was to reduce PoW rewards by 80% once the FFG layer was added. The new design is a bit different and I think it relies less on the PoW layer, but I'm not sure how it all fits together.

There's also Casper CBC, which is a more purist form of PoS they hope to implement later. That seems to be making good progress but it's still in the research phase.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#273

Post by jillesvangurp »

seibelj wrote: post #251
jillesvangurp wrote: post #220

I'm glad we switched away from Ethereum in favor of Stellar; it will take most of this year and probably next for the dust to settle on this. Until then, Ethereum continues to be unusable for serious dapps (which are rare to begin with) that have expectations that exceed having reasonable transaction throughput ( Until then, this is a toy platform for wannabe ICOs, misc gambling applications, and not much else that I…

Stellar, like Ripple, is centralized. They run an append-only database that only they can approve transactions on. Ethereum is attempting to be far more decentralized, which is why it’s so difficult. Stellar may solve problems but the types of problems it solves are nothing like what Ethereum is doing. And smart contracts are indeed important for many applications, despite their flaws.

Not as black and white as that. There are multiple companies running validators and you can choose to run your own. The transactions go to the same shared blockchain. Most of the significant tokens on Stellar have their own preferred validators for transactions. Each validator can configure consensus for their ecosystem as well. IMHO this is a reasonable compromise and not that different from Ethereum proof of stake in combination with a high stakes.

Ethereum is a nice science project so far that may one day result in something usable and when that happens, I will look at it again. However, that is not right now and we are trying to run an actual business right now. Most of the smart contracts I've seen are convoluted and tend to solve things that Stellar can do out of the box with zero programming.

The smart thing seems to be to reuse smart contracts from Github and avoid writing your own and thus avoid taking the risk of introducing exploitable bugs right in the heart of your financial operation.

Not to say that they are useless, but I have no need for them with Stellar. That saved us a lot of time so far.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#274

Post by jpadkins »

How different is PoS and PoW in practice (for ASIC eligible blockchains)?

I need to devote X units of currency to get Y hashrate to gain 51%. This is a stake, needed to achieve a certain level of PoW.

One difference is PoS requires stake only from internal currency (you must buy/mine a stake in that currency), vs ASIC PoW you can use a foreign currency to buy your 'stake'. This might mean PoS is less volatile, or resistant to external attacks.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#275

Post by jsutton »

jki275 wrote: post #209
baq wrote: post #166

Earlier quoted context omitted.

thus, the code has changed and with it the law. i don't see any problem here.

It was changed to fix the chain retroactively. That's the problem. If the ledger isn't immutable, it has no value as a ledger. The majority of the ETH world doesn't believe that, which is why people like Charles Hoskinson left it.

The whole point of the network is that it's owned and driven by the community. The community decided it was prudent to fork. It's not an issue with decentralization. The majority of the ETH world clearly does believe that, because ETH is where the vast majority of the community has stuck with.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#276

Post by woah »

xorcist wrote: post #172
ascendantlogic wrote: post #118

Earlier quoted context omitted.

1. No one was forced to run the new client that backed out the DAO 2. The existing network with the stolen funds and the DAO lived on as Ethereum Classic There was no central entity mandating that everyone was required to run the "new" network. It was entirely up to miners and people running nodes to decide which network they supported. Further the decision to back out the DAO hack was done via vote so that's another…

> There was no central entity mandating that everyone was required to run the "new" network Is the choice to run a less valuable network really a choice? Remember that the foundation people gave themselves most of the coins that will ever exist before the network was even started, even if a lot was sold long before there was a word for ICOs. Only ~ 20% or so of the supply will ever be mined. It is not economically ra…

What? Of course it’s a choice. And if the majority of people had made that choice, it would be the more valuable chain.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#277

Post by ramoz »

jzymbaluk wrote: post #244
ramoz wrote: post #237

Is there a public blockchain that can be used solely as a ledger, vs having to integrate with the currency model?

Hyperledger is a series of projects sponsored by the Linux foundation, Hyperledger Fabric [1] is an example of a blockchain project that is solely a ledger and smart contract platform, that isn't focused on the currency use case. AFAIK there isn't a canonical public Fabric blockchain, it's mostly focused on companies that want to roll their own private/permissioned blockchain, but I'm sure there are projects that are…

Thanks, yea we've considered implementing our blockchain requirements in Fabric and making it readable through a public api. But this is really then a "centralized" implementation.

The currency models of public blockchains are a distractor from the products (would-be DAPPs) that we want to deliver, where an immutable ledger is still a selling point.

Is it even worth it to do blockchain at this point? Maybe if we make the nodes completely public? idk.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#278

Post by dnautics »

wpietri wrote: post #252
dnautics wrote: post #110

Earlier quoted context omitted.

I'm sympathetic to the idea that technology decreases the per unit cost of productivity output, but the whole point of stimulating economic growth (via factors like inflation) doesn't work if the stimulus doesn't outpace the improvement in efficiency. Technology is fundamentally deflationary. So if you believe that inflation (and growth) is necessary for social stability, you must incentivise consumption beyond techn…

I think we're talking about two different things: economic value and the price of currency. I agree that you don't want currency deflation. But over decades, I don't think that has much to do with the value to users of goods. I can think of a few obvious places where they split. One is where we increase quality. A well-trained chef doesn't use significantly more resources than a bad cook, but the result is much bette…

> I agree that you don't want currency deflation

https://krugman.blogs.nytimes.com/2010/02/13/the-case-for-hi...

> Yet when you have very low inflation, getting relative wages right would require that a significant number of workers take wage cuts. So having a somewhat higher inflation rate would lead to lower unemployment, not just temporarily, but on a sustained basis.

Just to be clear that we understand the policy here in blunt terms: The point of inflation is to cheat the working classes out of their income so that the ruling class can lay claim to high employment metrics. This is the price of currency stability.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#279

Post by jki275 »

jsutton wrote: post #275
jki275 wrote: post #209

Earlier quoted context omitted.

It was changed to fix the chain retroactively. That's the problem. If the ledger isn't immutable, it has no value as a ledger. The majority of the ETH world doesn't believe that, which is why people like Charles Hoskinson left it.

The whole point of the network is that it's owned and driven by the community. The community decided it was prudent to fork. It's not an issue with decentralization. The majority of the ETH world clearly does believe that, because ETH is where the vast majority of the community has stuck with.

The majority of the community believed in the fork, that's what I meant if I wasn't clear.

That belief is contra to the concept of an immutable decentralized ledger, which means ETH isn't canonically a cryptocurrency, an immutable ledger, or a decentralized ledger anymore. The upcoming (well, it's never going to happen) POS fork is going to centralize it further among the very small number of massive whales who control the ecosystem.

Re: Ethereum Plans to Cut Its Energy Consumption by 99 Percent

#280

Post by pavel_lishin »

woah wrote: post #276
xorcist wrote: post #172

Earlier quoted context omitted.

> There was no central entity mandating that everyone was required to run the "new" network Is the choice to run a less valuable network really a choice? Remember that the foundation people gave themselves most of the coins that will ever exist before the network was even started, even if a lot was sold long before there was a word for ICOs. Only ~ 20% or so of the supply will ever be mined. It is not economically ra…

What? Of course it’s a choice. And if the majority of people had made that choice, it would be the more valuable chain.

Yes. People literally voted with their wallets.
Post reply on HN