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

[spiccioli]

I just had another thought:
To stay as a single p2pool: How about every payout address has its own, adjusting difficulty?
You would just have to broadcast a lot of data/chains. But since you can compare/convert each miners difficulty to the total p2pool hashing power, payout should be easy to calculate?
Just a thought..

That was the thought above which prompted by "essay".

The largest constraint is that as average share time decreases the orphan rate increases.  Eventually you reach a point where the share time is so small that orphan rate because astronomical. 

For example right now share difficulty is ~560.  To have a share difficulty ~280 would require 5 second average share time which means a significant increase in orphan rate of pool. 

The "concepts" I outlined above (and no I didn't come up with them) are methods to "compartmentalize" the network so that you can have more shares per second without higher oprhan rates.  All 4 "solutions" essentially do the same thing.  They allow greater than 6 shares per minute without an increase in orphan rate.  That is the ultimately the problem to solve.

DeathAndTaxes,

first and foremost p2pool current code could be modified to not accept new peers when total hash-rate of pool reaches a predefined limit, like 1 TH, this would force new p2peers to create a new instance and would auto-limit current p2pool.

best regards.

spiccioli.

[DeathAndTaxes]

I was actually wondering whether it is possible for each P2Pool instance to adjust its own difficulty, so that both weak miners and strong miners took 10 seconds to solve a block. Essentially have each P2Pool server dynamically adjusts the difficulty for its own miners only, as opposed to having one difficulty for the overall pool. The payout would then be (total number of shares submitted) * (hash rate). Since total number of blocks submitted will be, on average, the same for both large miners and small ones (difference only depending on amount of time they were running), the payout will essentially be only dependent on the hash speed.

It isn't EACH miner solves a share in 10 seconds (shares not blocks).  It is the entire p2pool solves 1 share (collectively) in 10 seconds.

This will require the chain to carry more data, but my main concern is that hackers could spoof their mining power to be way lower than it actually is, and then steal from everyone else by submitting way more than one share every 10 seconds. I realize this sort of defeats the purpose of a "verifiable chain," too...  

As I explained above that isn't a problem.  The difficulty of the share is proof of how much work was completed.  There is no way a miner can fake that.  A difficulty 1000 shares is worth 1000x as much as a difficulty 1 share and someone looking for 1000 difficulty shares will only find 1/1000th as many as someone looking for 1 difficulty shares.  Making variable shares "cheatproof" is trivially easy.  A couple lines of code is all it takes.  


Cheating isn't an issue, the issue is there is a limit on how quickly the network can efficiently handle shares.  p2pool raises difficulty to ensure time between shares remains ~10 seconds.  Not 10 seconds per miner , or 10 second for the average miner but 10 seconds FOR THE ENTIRE NETWORK.  Variable difficulty shares would still be constrained by that global "compromise".

[Gabi]

270ghash/s now!

Time to update the thread name 

[Wandering Albatross]

http://btcstats.net/p2pool/

The active users/addr doesn't match blockexplorer?  Why is that? Due to RRD?

Or do I misunderstand blockexplorer?
http://blockexplorer.com/t/93krV7V9FR shows 210 but your stats page has a max of ~140 active users/addr


RRD is not good, it averages away perfectly good data. (But people seem to use it a lot, maybe not knowing this.)

[breel]

Just moved one right to p2pool, but:
- I don't know where the receiving addresses are specified. Is p2pool just getting them through rpc from bitcoind?
- although my miner (diablo) seems fine, I see this in p2pool :

P2Pool: 17525 shares in chain (9178 verified/17529 total) Peers: 10 (0 incoming)
 Local: 393MH/s in last 10.0 minutes Local dead on arrival: ~3.6% (1-13%) Expected time to share: 1.6 hours
 Shares: 0 (0 orphan, 0 dead) Stale rate: Efficiency: Current payout: 0.0000 BTC
 Pool: 262GH/s Stale rate: 8.9% Expected time to block: 6.3 hours

which makes me wonder if it actually works..

Any idea?
Thanks!

[DeathAndTaxes]

Just moved one right to p2pool, but:
- I don't know where the receiving addresses are specified. Is p2pool just getting them through rpc from bitcoind?
- although my miner (diablo) seems fine, I see this in p2pool :

P2Pool: 17525 shares in chain (9178 verified/17529 total) Peers: 10 (0 incoming)
 Local: 393MH/s in last 10.0 minutes Local dead on arrival: ~3.6% (1-13%) Expected time to share: 1.6 hours
 Shares: 0 (0 orphan, 0 dead) Stale rate: Efficiency: Current payout: 0.0000 BTC
 Pool: 262GH/s Stale rate: 8.9% Expected time to block: 6.3 hours

which makes me wonder if it actually works..

Any idea?
Thanks!

1 p2pool share right now = >500 "normal" pool shares.  When it says expected time for 1 share is 1.6 hours it isn't kidding.  Due to variance it could be 2 minutes or it could be 8 hours.  1.6 hours is just the average for 1 share.  So seeing 0 shares isn't unusual.

As for your payment address yes p2pool gets it via RPC.  If you want you can force it to use a specific address by using command line option.

[kano]

Heh he beat me to it - but I will add, a variance of 8 times is not that uncommon on sha256(sha256())
So Maybe after 12 or 13 hours if you haven't found anything you might consider getting worried, but at least until then, that's not unexpected.

[twmz]

http://btcstats.net/p2pool/

The active users/addr doesn't match blockexplorer?  Why is that? Due to RRD?

Or do I misunderstand blockexplorer?
http://blockexplorer.com/t/93krV7V9FR shows 210 but your stats page has a max of ~140 active users/addr


RRD is not good, it averages away perfectly good data. (But people seem to use it a lot, maybe not knowing this.)

This site does not use RRD.  It takes the output of http://myp2poolserver:9332/users and parses it to determine the number of users.  It does this once every 5 minutes.  That said, it could be broken.  I'll review it tonight to see if it is counting wrong.

[breel]

@DeathAndTaxes & kano:
Thanks guys for the quick reply!

[breel]

Actually how do I use the "--address" option with merged mining?
Do I use a "," to separate them?

Thanks!

[Red Emerald]

Actually how do I use the "--address" option with merged mining?
Do I use a "," to separate them?

Thanks!

You can't specify an address for the merged chain (yet).  The merged mining still needs some work.  The address is currently fetched automatically via the RPC

[twmz]

http://btcstats.net/p2pool/

The active users/addr doesn't match blockexplorer?  Why is that? Due to RRD?

Or do I misunderstand blockexplorer?
http://blockexplorer.com/t/93krV7V9FR shows 210 but your stats page has a max of ~140 active users/addr


RRD is not good, it averages away perfectly good data. (But people seem to use it a lot, maybe not knowing this.)

This site does not use RRD.  It takes the output of http://myp2poolserver:9332/users and parses it to determine the number of users.  It does this once every 5 minutes.  That said, it could be broken.  I'll review it tonight to see if it is counting wrong.

I confirmed that everything is working as designed but that the design was flawed.  The number of users in the graph accuratly replicates what p2pool reports.  The problem with using that statistic is that the http://localhost:9332/users page only shows people that have submitted at least 1 share in the past 2 hours.  Small miners may not always find a share that often, and so aren't getting recognized.  I'll have to consider alternate approaches to approximating the number of users...

Stay tuned...

[forrestv]

I created this Google group for notifications about required updates and critical problems: http://groups.google.com/group/p2pool-notifications Subscribe if you wish; it's easier than keeping up to date with the forum thread!

[Rubberduckie]

using the newest version: http://u.forre.st/u/aswpssxw/p2pool_win32_7b6618c.zip

I get loads of rejects. (i'm using cgminer 2.2.3 with -I 8 -g 1 flags).

Previous version of P2Pool and cgminer 2.2.3 i was getting very few rejects and only 4% orphans, any ideas?


Edit:Updated to cgminer 2.2.5 and all good

[thatonegirlt]

This thread is massive and I can't find my problem by searching so here goes..

I used the n00b-guide plus the read me file to set up the pool and from cgminer sometimes it says Accepted and sometimes it says Rejected but Rejected much more than Accepted it appears.

The run p2p window says all the information like my payout address, the hashrate and estimated time for new block of the pool and my local hash rate at 0.0. The error I keep getting says..

Error binding to P2P port.  [Errno 10048]. Only one usage of each  socket address (protocol/network address/port) is normally permitted. Retrying in 3 seconds...

And it just keeps doing that indefinitely with the other messages throughout.

[Ente]

The error I keep getting says..

Error binding to P2P port.  [Errno 10048]. Only one usage of each  socket address (protocol/network address/port) is normally permitted. Retrying in 3 seconds...

You may have two instances of p2pool running. Accidental doubleclick, or one instance was in the autostart or something like that? A reboot should solve it for sure.
Or, the port is blocked by some firewall, which strangely reserves all ports for itself. Never heard of one like that.. Or you simply have not enough rights to open that port?
Which operation system, are you working with admin rights (if windows)?

Ente

[thatonegirlt]

I'm on Windows 7 64-bit. And I made sure I didn't have double instances running right now. Curious... I am at a college campus and they may be blocking the port somehow. I guess I'll have to check if I still get the error if I'm somewhere not on campus and report back.

[ancow]

I'm on Windows 7 64-bit. And I made sure I didn't have double instances running right now. Curious... I am at a college campus and they may be blocking the port somehow. I guess I'll have to check if I still get the error if I'm somewhere not on campus and report back.

P2pool is binding to a port on your machine. That means your college can't influence that, all they can influence is communication between your machine and the rest of the world.
The only way your college could influence p2pool's ability to bind to a local port is by having some software installed on your machine, and then it likely wouldn't make any difference where you are, it will always get blocked.

[Vanderbleek]

I'm on Windows 7 64-bit. And I made sure I didn't have double instances running right now. Curious... I am at a college campus and they may be blocking the port somehow. I guess I'll have to check if I still get the error if I'm somewhere not on campus and report back.

Do you have Litecoin running? IIRC, they use the same port by default. You'll need to tell either p2pool or Litecoin to use a different port, if that's the case.

[kano]

I'm on Windows 7 64-bit. And I made sure I didn't have double instances running right now. Curious... I am at a college campus and they may be blocking the port somehow. I guess I'll have to check if I still get the error if I'm somewhere not on campus and report back.

P2pool is binding to a port on your machine. That means your college can't influence that, all they can influence is communication between your machine and the rest of the world.
The only way your college could influence p2pool's ability to bind to a local port is by having some software installed on your machine, and then it likely wouldn't make any difference where you are, it will always get blocked.

Hm - I'd expect P2Pool to have problems if bitcoin is unable to access the internet ...

Also, it's not that unthinkable for environments (work, university, ?) to require certain software installed and running before you are able to access the internet - and that software could also cause issues.

No point assuming stuff until you know for sure it's true - though it is easy to check.