[1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool

[windpath]

A few weeks ago Petamine was thinking of joining P2Pool. They have about 1.15 Petahash or thereabouts. At the time P2Pool was barely hitting 1PH itself, so they would immediately have been over 51%. The pool dropped to something like 300-400TH/s for a bit after this, too.

They didn't join, but what could we have done if they did? A lot of people were saying "Join p2pool!", but then had they actually joined, people would have been saying "Don't join p2pool!!". It's not really a solution so much as an inevitable and hilarious outcome.

If I understand it correctly, it's a big deal solving this problem, because any solution that could solve it for p2pool could likely also be applied to the blockchain.

I would be happy to welcome PETAMINE to p2pool, I even offered to build them a custom front end.

The 51% on p2pool would not be an issue.

All that means is that they would get to pick the criteria for transactions included in 51% of the blocks found by the pool. If I'm PETAMINE I'm going to want the juicy transactions just like any other miner...

Right now we are barely 1% of the Bitcoin network, manipulating found blocks by p2pool would be completely irrelevant as far as bitcoin itself is concerned...

Also, They would NOT be able to influence payments for found shares, or alter a block that had been worked on by anyone outside their mine at all for that matter...

The only exception would be to withhold a found block and influence our stats, by withholding the block no one would get paid and our stats might crap out, but that is certainly not in their best interest...

[eleuthria]

If Peta-Mine was an even larger portion of p2pool, it could become a problem.  You don't have to be malicious to get more than your fair share.  Just like with Bitcoin, there's an inherent advantage to being a larger share of the network (or larger share of p2pool in this case).  The bigger you are, the more likely you are to confirm your own block/share in an orphan race.

Ironically, this would also reward them for having a *worse* connection to the rest of the p2pool network, because they're more likely to cause orphan races that way, and they would have a better than 50% chance of winning the race each time.  It requires no effort or malicious intentions, it's simply how the system would work in that scenario.

I actually made this argument back in early 2013 when people kept screaming that ASICMINER (at the time > 25% of the entire BTC network) should use p2pool.  Had they done it, they would have likely performed an "accidental 51%" attack because of their poor latency (being based in China) and the fact that they would have completely dwarfed the rest of p2pool's hashrate at the time.


It wouldn't be quite as big of a problem these days since p2pool increased the average share time.  It's still a slight issue where somebody with a significant hash rate (> 50%) is able to get a little more than their fair share, but it would have been extremely bad if p2pool still used the 10-second target time for shares.

[windpath]

 The bigger you are, the more likely you are to confirm your own block/share in an orphan race.

This is a valid point, but I would offer 2 counterpoints:

1. How often could this really occur, there are hundred(s) of p2pool nodes, the best connected are more likley to have the longest chain at 30 second shares, and there is only a 72 hour share chain window to play around with before it's vanished to the either... I imagine it could occasionally accidentally happen, but we are not strangers to orphaned shares as it stands

2. A big operation joining the pool, let's say 1PH/s, would cut our variance by more then 50% from were we are now, this would attract other miners to the pool who were held back by < 1 block per day... Those other miners would most likely remove a 51% advantage pretty quickly...

Again, you are right, but I just don't see it as a big threat, in fact I would welcome and encourage large mines to join p2pool!

Just my BTC0.02...

[eleuthria]

The bigger you are, the more likely you are to confirm your own block/share in an orphan race.

This is a valid point, but I would offer 2 counterpoints:

1. How often could this really occur, there are hundred(s) of p2pool nodes, the best connected are more likley to have the longest chain at 30 second shares, and there is only a 72 hour share chain window to play around with before it's vanished to the either... I imagine it could occasionally accidentally happen, but we are not strangers to orphaned shares as it stands

2. A big operation joining the pool, let's say 1PH/s, would cut our variance by more then 50% from were we are now, this would attract other miners to the pool who were held back by < 1 block per day... Those other miners would most likely remove a 51% advantage pretty quickly...

Again, you are right, but I just don't see it as a big threat, in fact I would welcome and encourage large mines to join p2pool!

Just my BTC0.02...

Don't get me wrong, in this case I don't think there is much threat of the "accidental 51%", or even much chance of any noticeable edge towards PetaMine, though there would be a very slight one.  It's just funny that like I posted earlier, how much of an advantage actually increases the worse their connectivity to p2pool is (assuming fully benign operators).

[-ck]

You still have the catch 22 problem if you bring in big mining operations like petamine: The bigger the p2pool hashrate, the harder it becomes for small miners. In precisely the opposite way from regular pools, the bigger p2pool is, the more it starts resembling solo mining for every miner. If the entire bitcoin network mined on p2pool, it would be like solo mining at 1/20th of the network hashrate (so like mining at a diff of almost 1 billion).

[errkal]

I have a question about fees.

Basically I want to know how they work.

I have a pool that has been running for a few days and 2 payouts have gone out of 0.002 each to the wallet address that is my BTC address.

I have a set a 5% pool fee as a maintenance fund, but nothing has gone to the p2pool address, how do fees actually work? I can seem to find anything documentation wise that actually explains how it works and how the payments are made, or more importantly when the payments are made.

I haven't set a p2pool address in the config in p2pool but it has picked one up from bitcoin as it connected using bitcoin.conf. Do I need to manaully say an address to use? in the p2pool config or is this address ok?

Thanks.

[windpath]

...how much of an advantage actually increases the worse their connectivity to p2pool is (assuming fully benign operators).

This is an interesting one, had to give it some thought... My assumption at his point would be that any latency to p2pool would also be reflected in bitcoind and perhaps remove any advantage they might have gained by causing orphan race losses for themselves on actual found blocks?

Also, its notable that there are a bunch of active nodes in China now, however this still might work in South America or Africa (Rio or Cape Town?) as I have not seen nodes there yet...

You still have the catch 22 problem if you bring in big mining operations like petamine: The bigger the p2pool hashrate, the harder it becomes for small miners. In precisely the opposite way from regular pools, the bigger p2pool is, the more it starts resembling solo mining for every miner. If the entire bitcoin network mined on p2pool, it would be like solo mining at 1/20th of the network hashrate (so like mining at a diff of almost 1 billion).

Agreed, but again, this is a problem I'd love for p2pool to actually have (on a bigger scale)...

The lower threshold for getting a payout in ~most~ blocks seems to be around 40 GH/s right now. Would bumping that to 80-100 GH/s by adding 1 PH/s to the network impact that many miners? It looks like under 2% of pool miners right now...

It is a very real challenge/problem, but I don't think it impacts that many miners, and the ones it is currently impacting would really not be contributing that much to the pool to begin with (I know poor excuse, I'd like to include everyone...).

Actually making the problem worse (by growing the pool speed and increasing the minimum effective hash rate per-miner) might also prompt more active work on a solution... Or slower miners could join together and simply create an alternate share-chain (I don't think this is the best long term solution, but it would work today...)

Again, agree its a real problem, but one I'd love for us to have it on a larger scale so that it might "force" a solution...

[windpath]

I have a question about fees.

Basically I want to know how they work.

I have a pool that has been running for a few days and 2 payouts have gone out of 0.002 each to the wallet address that is my BTC address.

I have a set a 5% pool fee as a maintenance fund, but nothing has gone to the p2pool address, how do fees actually work? I can seem to find anything documentation wise that actually explains how it works and how the payments are made, or more importantly when the payments are made.

I haven't set a p2pool address in the config in p2pool but it has picked one up from bitcoin as it connected using bitcoin.conf. Do I need to manaully say an address to use? in the p2pool config or is this address ok?

Thanks.

There are 2 types of fees you can set, one for forrestv (--give-author), and one for an address you designate (-f). If you do not set -f it will pull an address from bitcoind, this is not a good practice in my opinion as you don't want to keep any Bitcoin on a public p2pool node for security reasons.

Both fees work the same way, they are a % of found shares by the node, not an actual % of earnings... So in your case 5% of found shares would go to the address you designate with -f.

[mdude77]

The lower threshold for getting a payout in ~most~ blocks seems to be around 40 GH/s right now. Would bumping that to 80-100 GH/s by adding 1 PH/s to the network impact that many miners? It looks like under 2% of pool miners right now...

I'm curious how you came up with 40 GH/s?  I have one S1 (180 GH/s) that I'm not putting on p2pool because the variance is too much for me.  If my S2s worker properly, I'd put them all there, but they don't, so the S1 stays with my S2s.

M

[windpath]

The lower threshold for getting a payout in ~most~ blocks seems to be around 40 GH/s right now. Would bumping that to 80-100 GH/s by adding 1 PH/s to the network impact that many miners? It looks like under 2% of pool miners right now...

I'm curious how you came up with 40 GH/s?  I have one S1 (180 GH/s) that I'm not putting on p2pool because the variance is too much for me.  If my S2s worker properly, I'd put them all there, but they don't, so the S1 stays with my S2s.

M

http://minefast.coincadence.com/p2pool-stats.php

Click on miners tab and sort by hash rate, you can look at miner dashboard for any of them and see payouts...

[mdude77]

The lower threshold for getting a payout in ~most~ blocks seems to be around 40 GH/s right now. Would bumping that to 80-100 GH/s by adding 1 PH/s to the network impact that many miners? It looks like under 2% of pool miners right now...

I'm curious how you came up with 40 GH/s?  I have one S1 (180 GH/s) that I'm not putting on p2pool because the variance is too much for me.  If my S2s worker properly, I'd put them all there, but they don't, so the S1 stays with my S2s.

M

http://minefast.coincadence.com/p2pool-stats.php

Click on miners tab and sort by hash rate, you can look at miner dashboard for any of them and see payouts...

I know I'll get a payout.. but variance is a bear.

M

[windpath]

The lower threshold for getting a payout in ~most~ blocks seems to be around 40 GH/s right now. Would bumping that to 80-100 GH/s by adding 1 PH/s to the network impact that many miners? It looks like under 2% of pool miners right now...

I'm curious how you came up with 40 GH/s?  I have one S1 (180 GH/s) that I'm not putting on p2pool because the variance is too much for me.  If my S2s worker properly, I'd put them all there, but they don't, so the S1 stays with my S2s.

M

http://minefast.coincadence.com/p2pool-stats.php

Click on miners tab and sort by hash rate, you can look at miner dashboard for any of them and see payouts...

I know I'll get a payout.. but variance is a bear.

M

It can be... I bought 2 S1s close to the end at .5, had a PSU, they have made ROI on p2pool:

http://minefast.coincadence.com/miner.php?id=13AxQAKFiy6Dks5Jwdbah3rhMJAJ9UzwHk

Obviously this does not take power costs into consideration, but plenty of time left running in the black...

[sconklin321]

I have a question about fees.

Basically I want to know how they work.

I have a pool that has been running for a few days and 2 payouts have gone out of 0.002 each to the wallet address that is my BTC address.

I have a set a 5% pool fee as a maintenance fund, but nothing has gone to the p2pool address, how do fees actually work? I can seem to find anything documentation wise that actually explains how it works and how the payments are made, or more importantly when the payments are made.

I haven't set a p2pool address in the config in p2pool but it has picked one up from bitcoin as it connected using bitcoin.conf. Do I need to manaully say an address to use? in the p2pool config or is this address ok?

Thanks.

There are 2 types of fees you can set, one for forrestv (--give-author), and one for an address you designate (-f). If you do not set -f it will pull an address from bitcoind, this is not a good practice in my opinion as you don't want to keep any Bitcoin on a public p2pool node for security reasons.

Both fees work the same way, they are a % of found shares by the node, not an actual % of earnings... So in your case 5% of found shares would go to the address you designate with -f.

It's actually -a to set the node payout address, -f is the fee %.

[windpath]

It's actually -a to set the node payout address, -f is the fee %.

That is correct, good catch, thanks!

[errkal]

Ahhh ok that makes sense, so if I found 100 shares 5 of them would go to the pool

[windpath]

Ahhh ok that makes sense, so if I found 100 shares 5 of them would go to the pool

About 5... Also subject to variance so it not always exact...

[errkal]

Good old variance

[raskul]

iceland node now available on www.centralcavern.uk

http://is.centralcavern.uk:9332/static/

[sconklin321]

Wow did someone big just join the network?  We're up to almost a Petahash currently.

Edit: Actually broke it while I was writing this.

[bryonp]

Wow did someone big just join the network?  We're up to almost a Petahash currently.

I was just going to say... it has been jumping from 800th to 500 th back to 900 th down to 450 th and now
1 PH...?

And my payouts are dropping big time