5 ms·

I haven't been keeping up with the Eth / PoS development, do we have any preliminary numbers on how much energy this switch is hoped to save? Same goes with sh

by olodus 6y ago

I haven't been keeping up with the Eth / PoS development, do we have any preliminary numbers on how much energy this switch is hoped to save? Same goes with sharing, how much is it argued this will speed up the transaction process?

rglullis 6y ago
Validators will need to run a computer with reasonably modern specs (at least 1TB SSD, 8-16GB RAM, multiple cores) in order to collect fees and not have their funds slashed. IIRC you need to have at least 70% uptime to not lose any funds, and to be a good investment you need 99%+ uptime. If you consider that modern systems with the required specs can run on 10W, you should get ~90kWh in a year. Transaction-speed wise, there is a lot of talk about 100000+ TPS, but the general consensus seem to be that this will only be achieved with extra L2 support and that a more realistic number for "pure on-chain" throughput is in the high-hundreds / low thousands.
- diang 6y ago
  Competing with Bitcoin's TPS in the single digits, high-hundreds/low thousands is a big leap... but still not enough to replace the higher order global currencies. In order to gain the necessary L2 support, they'll have to establish dominance relative to bitcoin with an equally lacking TPS. Seems like a "If they come, we will build it" problem.
  - rglullis 6y ago
    Plenty of L2 solutions already exist and can be used with Ethereum already. If your single comparison point is the existing banking and credit card payment networks, you can take a look at e.g, Raiden which is already deployed on mainnet and can scale with the number of the nodes participating in the network. Hundreds of thousands of payment transactions per second are already possible. The main issue is that most of them still are very complex to get fully operational, there is no big consensus among users on "what-to-use-when" and they all require a non-insignificant cost to setup in terms of all the contract transactions needed to get in and out of the chain. There is no reason to doubt that they be more mature along with ETH2 and that that L1-L2 will complement each other nicely.
    - diang 6y ago
      I could take a look at e.g, anything... but when you add a caveat of participation, scaling to infinity is inherently possible. If we all chose to do something, we could all do it. The problem comes with attracting such participation, which isn't warranted without that very participation having already been established. Carts and horses and such.
      - rglullis 6y ago
        First you said it was a "if they come we will build it" problem, I just showed you one example that demonstrates that your claim is false. Also, I said that there are many things yet in the way to facilitate adoption, no need to move the goal-posts.
        diang 6y ago
        You don't seem to understand what my claim is, and your example was just speculation that doesn't reflect reality. I never said there were goal-posts... if anything I said they couldn't exist, and even if you could score, you'd get points in the wrong game.
        rglullis 6y ago
        If I understand you, you are saying that Ethereum is not important because people are not valuing as highly as BTC, so the transaction throughput is irrelevant and L2 is not going to be needed. This is akin to saying that it doesn't matter that e.g Brazil has a more advanced banking retail industry compared to the arcane networks from the US because the Brazilian economy is not as big as the US. This is silly, because not only the two economies can co-exist, there is no technical aspects that stops other chains to adopt PoS as well it also ignores that the value can be migrated from one chain to another. Increasing throughput in Ethereum would make everything more valuable. Just as an example, people can get their BTC and transform wrap them to use in the ethereum chain. With ETH2, it will be cheaper to wrap BTC and cheaper to use it in transactions.Its utility will be increased.
  - CryptoPunk 6y ago
    >>In order to gain the necessary L2 support, they'll have to establish dominance relative to bitcoin with an equally lacking TPS. In terms of utilization, Ethereum already has, exceeding Bitcoin in transactions per day: https://etherscan.io/chart/tx https://etherscan.io/chart/tx And doing so while the average complexity of transactions, which is measured by the number of gas units a transaction consumes, has increased: https://etherscan.io/chart/gasused https://etherscan.io/chart/gasused It has also overtaken Bitcoin in the economic demand for its blockspace, with Ethereum miners earning more in fees than Bitcoin miners: https://coinmetrics.io/charts/#assets=btc,eth_log=false_left=FeeTotUSD_zoom=1590364800000,1598140800000 https://coinmetrics.io/charts/#assets=btc,eth_log=false_left...
drcode 6y ago
Once the transition is complete, I would expect the energy counsumed to be only 0.1% of a POW chain, so a thousandth the amount of power. (but I haven't attempted to work out an exact number and am not an expert on these matters)
leesalminen 6y ago
One of the main drivers behind ETH2 is sharding, which will allow for 10,000s txn/sec.
- ishishkin 6y ago
  But at what cost? There is no chance it could preserve the same level of decentralization.

by olodus 6y ago

rglullis 6y ago

Validators will need to run a computer with reasonably modern specs (at least 1TB SSD, 8-16GB RAM, multiple cores) in order to collect fees and not have their funds slashed. IIRC you need to have at least 70% uptime to not lose any funds, and to be a good investment you need 99%+ uptime. If you consider that modern systems with the required specs can run on 10W, you should get ~90kWh in a year. Transaction-speed wise, there is a lot of talk about 100000+ TPS, but the general consensus seem to be that this will only be achieved with extra L2 support and that a more realistic number for "pure on-chain" throughput is in the high-hundreds / low thousands.

diang 6y ago
Competing with Bitcoin's TPS in the single digits, high-hundreds/low thousands is a big leap... but still not enough to replace the higher order global currencies. In order to gain the necessary L2 support, they'll have to establish dominance relative to bitcoin with an equally lacking TPS. Seems like a "If they come, we will build it" problem.
- rglullis 6y ago
  Plenty of L2 solutions already exist and can be used with Ethereum already. If your single comparison point is the existing banking and credit card payment networks, you can take a look at e.g, Raiden which is already deployed on mainnet and can scale with the number of the nodes participating in the network. Hundreds of thousands of payment transactions per second are already possible. The main issue is that most of them still are very complex to get fully operational, there is no big consensus among users on "what-to-use-when" and they all require a non-insignificant cost to setup in terms of all the contract transactions needed to get in and out of the chain. There is no reason to doubt that they be more mature along with ETH2 and that that L1-L2 will complement each other nicely.
  - diang 6y ago
    I could take a look at e.g, anything... but when you add a caveat of participation, scaling to infinity is inherently possible. If we all chose to do something, we could all do it. The problem comes with attracting such participation, which isn't warranted without that very participation having already been established. Carts and horses and such.
    - rglullis 6y ago
      First you said it was a "if they come we will build it" problem, I just showed you one example that demonstrates that your claim is false. Also, I said that there are many things yet in the way to facilitate adoption, no need to move the goal-posts.
      - diang 6y ago
        You don't seem to understand what my claim is, and your example was just speculation that doesn't reflect reality. I never said there were goal-posts... if anything I said they couldn't exist, and even if you could score, you'd get points in the wrong game.
        rglullis 6y ago
        If I understand you, you are saying that Ethereum is not important because people are not valuing as highly as BTC, so the transaction throughput is irrelevant and L2 is not going to be needed. This is akin to saying that it doesn't matter that e.g Brazil has a more advanced banking retail industry compared to the arcane networks from the US because the Brazilian economy is not as big as the US. This is silly, because not only the two economies can co-exist, there is no technical aspects that stops other chains to adopt PoS as well it also ignores that the value can be migrated from one chain to another. Increasing throughput in Ethereum would make everything more valuable. Just as an example, people can get their BTC and transform wrap them to use in the ethereum chain. With ETH2, it will be cheaper to wrap BTC and cheaper to use it in transactions.Its utility will be increased.
- CryptoPunk 6y ago
  >>In order to gain the necessary L2 support, they'll have to establish dominance relative to bitcoin with an equally lacking TPS. In terms of utilization, Ethereum already has, exceeding Bitcoin in transactions per day: https://etherscan.io/chart/tx https://etherscan.io/chart/tx And doing so while the average complexity of transactions, which is measured by the number of gas units a transaction consumes, has increased: https://etherscan.io/chart/gasused https://etherscan.io/chart/gasused It has also overtaken Bitcoin in the economic demand for its blockspace, with Ethereum miners earning more in fees than Bitcoin miners: https://coinmetrics.io/charts/#assets=btc,eth_log=false_left=FeeTotUSD_zoom=1590364800000,1598140800000 https://coinmetrics.io/charts/#assets=btc,eth_log=false_left...

drcode 6y ago

Once the transition is complete, I would expect the energy counsumed to be only 0.1% of a POW chain, so a thousandth the amount of power. (but I haven't attempted to work out an exact number and am not an expert on these matters)

leesalminen 6y ago

One of the main drivers behind ETH2 is sharding, which will allow for 10,000s txn/sec.

ishishkin 6y ago
But at what cost? There is no chance it could preserve the same level of decentralization.