I recently switched from running bitcoind to btcd on linux.

Once golang was installed it’s pretty much an automated installation process – 1 command to auto download and install the btcd package. A well-documented sample config file is provided that was easy to tailor. Launched it and done.

Full network sync from scratch took 26 hours. Memory usage is in the 200MB range. Couple of days in it’s been rock solid and seems to be working great.

I’m hugely impressed with what the devs have achieved. Their code appears to be clean, modular, well documented and comprehensively covered by tests. So a great base to build on.

Maybe a re-write could also be an opportunity to tinker with the protocol to improve things. Not sure if these are totally practical but a couple of thoughts I had:

At the moment, as I understand it, there is a single difficulty for the sharechain. I was penciling out a potential alternative whereby each node attempts to target 10 shares in the PPNS sharechain. This would be big enough to minimize volatility but small enough for there to be room in the PPNS for more lower hashrate miners.

Fixing the impact of the 30 second share time causing difficulties with some ASIC miners would good too. Maybe one possibility is that the last 20 shares in the sharechain are not paid. So miners would work on the same blockheader for 10 minutes not updating it for each and every share found. Then when the 20th new share is found a new blockheader is created that incorporates them and the process starts again. I think this would balance out over time to be fair and maybe fix some of the miner compatibility issues?

I was thinking about having shot at implementing p2pool in Go. But it's a big job and complicated. I don't think I have the time or skill to do it.

Thanks, any issues? Or is it a drop in replacement once compiled

I'm guessing you would need to run the original p2pool once the C++ relay is running?

Maybe consider running one or more p2pool nodes for people to connect to? You set the fee to whatever you want (although 0% is common).

I have a couple of publicly available nodes:

london.bitroute.io:9332
newark.bitroute.io:9332

and there are loads more available: http://p2pool-nodes.info/

If you're wanted to run a more conventional pool I think it would be hard to generate sufficient hashrate to reduce volatility to an acceptable level.

It's good to see something happening on development front.

If they attach a few PH/s to could this swamp p2pool though? Won't this cause share difficulty to rise and push out smaller miners even more quickly than at present?

I had been thinking about how to minimize the impact of wide ranges of hashrates to help smaller miners reduce their volatility. Like enforcing variable difficulty levels to target a small number of shares so there is room for more individual miners in the PPNS payout.

Maybe Bitmain have thought about all this and are incorporated something to preserve the p2p nature when you have one node (or a few collected ones) which are much bigger than others.

Will be interesting to see what they are up to.

Having a chance to read a relatively compact yet complex and well structured bit of C is a great resource for beginners to learn from an expert. So thanks a lot for creating and open-sourcing the code.

Could you recommend a good resource to learn modern style, multi-threaded and safe C? I've got K&R but that doesn't have the paradigms and approaches that I think would actually be used these days - or should I just go to kernel.org and start reading


On the pool, it's great to give another option for miners to choose from. As smaller miners find that it's not economic to mine with low volatility, perhaps embracing the high volatility and having a lottery ticket option can give a good reason for some of these guys to keep going.

Good luck - I hope someone finds a block or two on the pool soon.

Basic question but do we know why some hardware isn't working well with p2pool?

I assume the hardware is driven by the same mining software like cgminer? Which is using the same protocols to talk to p2pool (stratum?) as GHash.

So is it something to do with polling frequency / latency that is p2pools fault or what?

My take was that what Matt is doing is very cool and when blocks are several MB in size something like this or Gavin's alternative proposal for eliminating retransmission of transactions with every block will be important.

For me, each of my nodes has a connection which p2pool and bitcoind never come close to utilising. So I didn't think the block size was a particularly big deal at the moment from a speed perspective. Don't know if that's been actively researched.

If Matt's built it and it works then I see no reason not to connect to one of his nodes to gain the benefit of it.

London  | london.bitroute.io:9332 | 0% | Bitcoin
Newark | newark.bitroute.io:9332 | 0% | Bitcoin

Some great thoughts windpath.

What stops someone taking your code, running the complex node and replacing your advertisements with their own? Does this just create a way for others to monetise your work?

Or would the complex node stuff serving the API stay closed source?

They wouldn't need to resubmit to the main repo. They just fork it (like you did) include all the new development work done and paid for but leave out the little bit that makes the donation and then run the node. Presto - 0% fee. They could either release it publicly if they wanted to be an ass or better still just do it on the quiet and leach. It's just the existing system but without the convenient command line option that forrestv put in.

But the block chain concept already solved this problem. The same reason a miner can't just give themselves 50 BTC when they mine a block or double spend an input (analogous to removing a development charge). Although any miner could create a block awarding themselves extra BTC the rest of the bitcoin network agrees by consensus that this new block is not valid and simply ignores it - it doesn't enter the block chain. The rules of the system are encoded by the majority consensus of miners. That's the power of the 51% attack. With 51% you can change whatever rules you want and benefit yourself.

Taking this to the equivalent situation in the p2pool share chain. If the majority of miners on p2pool agree (or rather their nodes enforce) that a valid share must include a 1% donation to whomever then bang it's the law. Doesn't matter what an individual does with his nod. Nobody else will recognise it so he doesn't get that share. It's a hard fork though because you effectively create an entirely new p2pool which has an incompatible rule with the original p2pool.

So the technology is there to do something like that. But trouble is perhaps it's not really p2p anymore. You've just invented a semi-centralised pool that enforces a payment to the boss - even a benevolent one. So politically this could be hard to achieve in practice.

I agree with everything you said KGB. But given its open source what stops someone from stripping out the mandatory fee and gaining the benefits without paying.

The share chain solution is that if the majority of p2pool are willing to play by these rules and reject freeloading shares that don't include the development fee then it would be enforced. That would be a hard fork so would need managing.

Will the majority of miners play ball though?

I just posted this on another thread but it's relevant so what the heck:

What the banks are truly terrified of is money laundering and sanctions. Several major banks have been fined many billions of dollars this year over not doing what they should in this area. As it stands the bitcoin world just isn't up to scratch in complying with these rules. If a bank facilitates it they are liable.

So my guess is that some banks have decided that the risk profile of any bitcoin business, especially ones that provide bitcoin to fiat exchange, are too risky to deal with.

Those that are well backed and have professional staff and controls that are proven to be good enough (maybe coinbase say) will get an exemption.

It seems harsh but honestly it's how the entire financial system deals with any money broking business. I've seen it first hand because I've consulted for companies and banks to help them comply with the rules.

As bitcoin matures and becomes a serious medium for wealth transfer this kind of thing is going to step up. In the big scheme of things it's maybe a good step toward decriminalizing the currency and helping mainstream adoption. But it's a shame if it's shuts down some entrepreneurialism as a result.

I don't think banks care about bitcoin as a competitive threat. The big banks don't make money with 'transaction' banking. Most of them lose money on this but do it because they have to. They make money on loans, product fees (like mortgage arrangement fees) and investment banking.

If bitcoin became a major currency they would just love to provide currency hedging and all their other products off the back of it.

But what they are truly terrified of is money laundering and sanctions. Several major banks have been fined many billions of dollars this year over not doing what they should in this area. As it stands the bitcoin world just isn't up to scratch in complying with these rules. If a bank facilitates it they are liable.

So my guess is that some banks have decided that the risk profile of any bitcoin business, especially ones that provide bitcoin to fiat exchange, are too risky to deal with.

Those that are well backed and have professional staff and controls that are proven to be good enough (maybe coinbase say) will get an exemption.

It seems harsh but honestly it's how the entire financial system deals with any money broking business. I've seen it first hand because I've consulted for companies and banks to help them comply with the rules.

Why do you think that? People have to pay their bills in fiat. So if you're working out what someone 'earned' from their work doesn't the exchange rate to USD matter?

I don't know but I imagine that back in 2011 Forrest made p2pool not for the money. But I think the world has changed since then. A lot of the modern miners spending $1,000s on new asics, the makers of those asics and the commercial pools attracting them are doing it primarily for the money.

Wish I had a solution.

I wasn't clear what I meant perhaps. I agree there is no way p2pool (or bitcoin) would work on a closed source basis. It has to be open for large number of reasons.

My 'problem' with open source for p2pool, for want of a better word, is that developers cannot enforce users of the software to pay for the development effort. So free as in speech and not free as in beer. While lots and lots of opensource projects are done for the love of the craft I'm not sure how that would work for something like p2pool.

I guess you could say the same about Bitcoin as a whole. But there are ways for developers supporting the bitcoin network to get paid something for their effort.


I wonder how much that is if you weight it by the bitcoin price at the time it was donated. Either way it's maybe not enough to expect someone with the necessary skill to put in the amount of effort required.

Hard to see a good solution if it's to be open source too. Any attempt by a developer to gain a reward for their efforts can be undone with a quick hack.

Solutions that require altruism from miners are possibly doomed to fail?

Equally I can't see why a talented coder would work hard to support a system making everyone else money except him.

Yeah, I read some of that. I saw the rate of development had tailed off and the issues with some mining hardware. Hadn't realised how much of an issue that was though. Will be interesting to see how the p2pool community deals with the lack of development. Would be a huge shame if it collapses.

Actually I was really tempted to try to reimplement some of p2pool in a systems language like go. Both as a bit of a fun project and also to learn the protocol. I couldn't find any documentation on the protocol beyond the python source though and it doesn't look the easiest to follow.


Windpath very graciously sent me a PM to advise that I was actually paying Forrestv 0.005% by accident instead of 0.5%. I need to see the config file to work out how I messed that up! But as it turns out it was a nil donation anyway - sorry about that

I think that for some people there isn't a net benefit to running a proper node. If they are running one at home then are probably hidden costs like less efficient mining and orphan shares which add up. For instance having ISP QOS filtering and firewalls or random middle of the night network maintenance downtime etc etc. Not to mention the potential security risks of opening up a home computer to the wider internet. You wouldn't believe how many Chinese hackers try to hack these servers every hour when I look at the firewall logs - scary stuff.

Where people give the service away for 0% fee like coincadence then that's going to be pretty hard to beat competitively but there will only be so many miners that can connect to that before the amount of money he's giving away will make it untenable. Good advertising for the excellent price ticker service though and a good way to support the community so I'm all for it.

In competitive spirit, seeing as the hashrate on my node is currently 0, I can offer a highly competitive fee reduction to 0.55% (by stiffing forrestv). I don't even have to change the config file!!

When I get a chance later today I'll set it to 0%. It's not really about the money so I'm happy to run the nodes for several months without any income.  I don't like the way things are going with GHash etc and without p2pool being a credible alternative for larger miners it will die out.

It would be great to get some evidence that a stable and well connected node can pay for itself in improved mining revenue over a homebrew solution.

Cheers
Paul

Thanks for the comments. I appreciate it.

Agree that a 1% total fee is something for people to think about. Especially when there are 0% nodes around.

The goal is that a high quality service would offset the costs by producing a superior return. Doesn't take much downtime or orphan shares to 'cost' 1% of your income. Of course my nodes are pretty new so there's no track record of that superior return of course to be fair 

In the overall picture of how volatile the mining business is (given diff changes, new asics, power costs, USD/BTC exchange rate) in a way a fixed 1% is relatively small to the overall risk/returns. I guess I'm hoping that there may be some miners out there who are willing to pay a moderate fee for a reliable good quality service. Maybe mining is too hard nosed for that though.

You've got a good point about development funding. Although, if nobody makes donations to pay for p2pool development then will it continue to exist? Then we're left with the commercial pools using proprietary software and charging 3% fees with all the centralization and other issues that go with it. Tragedy of the commons I guess.

If there really is no significant movement in p2pool development for a few months then I imagine the community will just end up cutting donations to zero (if they're not there already) and hope that someone steps in to maintain the code.

Let's see what happens anyway. If I can attract some hashrate to profile it I'd be interested to know if efficiency/returns are better than other 0% pools. If not then I guess we'll see.

Would love to know what other miners think. Will you only consider using a 0% pool?

Hi all

I have set out to make the best public p2pool node network out there. It's a big task to convince miners that it's worth moving their hard earned GHash to a new node/pool but here are the reasons why I hope you will consider using my bitroute.io nodes:

1. Mining using p2pool will return the highest ROI for your GHash - possibly over 100% fair value on a long run basis if you have the right node. But given the thread you probably know that.

2. Supporting p2pool is in the long term interests of a decentralised and healthy bit coin network. Again, telling you something you already know.

3. You need access to the lowest latency between your miner and the p2pool network to achieve over 100% efficiency.

Bitroute has two permanently connected p2pool and bitcoin servers - one in Newark, NJ, and one in London, UK, with a 75ms RTT ping between them and virtually no latency variance. Each node (a Linode 4) has a 500Mbit otherwise unutilised network pipe, 4Gb of RAM and 4 core processor.

So the time between your miner finding a share and it being widely broadcast to the global p2pool network will be as low as possible so you can get more than your fair share of shares.

There is also another full bitcoin node in Tokyo permanently connected to London and Newark to make sure that any blocks mined are immediately propagated to miners in Asia as well as Europe and the US to minimize the risk of orphaned blocks and lost revenue.

4. You need absolutely reliable nodes to minimize downtime. Bitroute uses production quality Linode servers using a highly stable Linux distribution and a locked down firewall.

They run the latest release (0.9.2.1) of Bitcoin Core and the latest version of p2pool from Github. There are no fancy settings, altcoins, weird configuration files or other stuff being run on these machines to interfere with their reliability or performance. As far as possible with current Bitcoin software these should be rock solid.

In addition each machine checks once a minute that the bitcoind and p2pool nodes are running. Should either bitcoind or p2pool crash unexpectedly then the server will automatically restart them.

5. The nodes will pay 0.5% of returns to Forrest, the developer of p2pool, to ensure he can continue to support and develop the software we rely on.

6. The nodes run a stock bitcoin configuration. No abusive block sizes or non-standard relaying policies etc. The aim is to mine and benefit the bitcoin network rather than try to game the system for an extra satoshi or two.
 

The downsides:

7. Bitroute will collect a 0.5% fee from miners to pay for the service.

8. As with all PPNS/p2pool nodes you'll need to mine for about a day until your proportion of payouts starts to reach its full potential.


How:

If you're interested please point your miner (with long polling connections if possible) to whichever of these machines is closest to you (they respond to a ping so feel free to measure):

newark.bitroute.io:9332

london.bitroute.io:9332


You can also see their individual stats by pointing your web browser at the addresses above as usual.


If you have any concerns or suggestions on what would make this p2pool node more appealing just please let me know.


Thanks
Paul