BITMAIN announces Antpool

[jonnybravo0311]

Actually, I can see how to pull off the "regular" pool on top of p2pool.  When p2pool mining, you must supply your BTC wallet address, and it must be valid.  If it isn't, your hash rate is directed towards the node's payout wallet.  Try it some time.  Start up your own node and connect to it with a user name of anything BUT a BTC address.

Now you just place a typical PPLNS framework on top of your node.  People register with a user name.  When they connect with that name, since it's not a BTC address, their hash rate gets added to the node's default address.  Now you have all of this hashing power on one address that is contributing to the share chain.  As soon as a block is found, that address is paid.  Then, you just divvy out the payment to the miners based on a standard PPLNS system.

So... I can see how it can be done, but it adds a centralized feeling to an inherently decentralized protocol.  In other words, for this magic to work, you simply cannot move from node to node as you can with a regular p2pool node.  The new node you move to / register with has no knowledge of any work you personally have contributed on another node.  Basically what I've described as a possible approach does nothing more than give people the ability to setup and run their own traditional pools using the p2pool network as a backend framework.  Sure, it solves the "small miner" issue, but at what cost?

[1713855800]

1713855800

[1713855800]

1713855800

[murraypaul]

Now you just place a typical PPLNS framework on top of your node.  People register with a user name.  When they connect with that name, since it's not a BTC address, their hash rate gets added to the node's default address.  Now you have all of this hashing power on one address that is contributing to the share chain.  As soon as a block is found, that address is paid.  Then, you just divvy out the payment to the miners based on a standard PPLNS system.

That does nothing to solve the lost work due to the short p2pool share time.

[jonnybravo0311]

Now you just place a typical PPLNS framework on top of your node.  People register with a user name.  When they connect with that name, since it's not a BTC address, their hash rate gets added to the node's default address.  Now you have all of this hashing power on one address that is contributing to the share chain.  As soon as a block is found, that address is paid.  Then, you just divvy out the payment to the miners based on a standard PPLNS system.

That does nothing to solve the lost work due to the short p2pool share time.

No, it doesn't, and I didn't state that this would address it.  All I stated was how you could put a traditional PPLNS payout framework on top of a p2pool backend.  The frequent restarts... now there's a problem that needs solving, and I'm not sure Bitmain has pulled it off.  Will certainly need to see what they're attempting if/when they actually do drop their source code in github.

[dogie]

Okay, there was some confusion over the word 'solo' when it was translated, it will get corrected in the announcement and site tomorrow but here is the gist of it.

What has been referred to as 'solo' (stratum.antpool.com:3333, or solo.antpool.com:3333, solo.antpool.com:25) is NOT individual miners trying to find ~25btc blocks. It is an independent pool run by Bitmain. The P2P side of things (p2p.antpool.com:3333 and p2p.antpool.com:25) is separate, and tacks on to the current p2p network [I don't have more information in regards to its technical set up as p2p has only just gone live].

tldr, solo = bitmain's pool, NOT p2p and NOT solo mining,
p2p = p2p

[windpath]

As far as I can see p2p.antpool.com:3333 looks like a standard node, with a single miner @ ~ 4TH/s, that has yet to even find a share?

Here is the node address payout history (or lack there of): http://minefast.coincadence.com/miner.php?id=1MjeEv3WDgycrEaaNeSESrWvRfkU6s81TX

Edit: Where is the latency based routing they talked about? I'm hitting from the US East coast and being routed to 182.92.226.212, in China.

[jonnybravo0311]

As far as I can see it looks like a standard node, with a single miner @ ~ 4TH/s, that has yet to even find a share?

Here is the node address payout history (or lack there of): http://minefast.coincadence.com/miner.php?id=1MjeEv3WDgycrEaaNeSESrWvRfkU6s81TX

Well, the node's been up for less than half an hour.  So much for their "extensive testing" claims.  Standard p2pool default front end.  Only thing different is they've exposed port 3333 instead of the typical 9332.

[IYFTech]

What?!  Is he for real?  

Wrong again dogie, wrong again........

So I'm an official shill in his book now am I? Cool.

Edit: Wait, why hasn't he left me any? That's not fair - I want some.........

[norgan]

They aren't providing a charity as they are pocketing the transaction fees. What they appear to be doing is providing an agent as such. A mechanism whereby you mine on their node to the default address that then appears as a larger miner. Since its larger there is less variance. They are then calculating the shares to miners on the node and paying them from that nodes wallet.
It is centralizing the smaller miners but still hashing on p2pool's sharechain.
It's a novel approach to trying to solve variance and similar to having a secondary sharechain. It will be interesting to see the code and how it works out the subchain payments.
I must say it does sound like what Kano and ck have been doing, like a ckpool instance connecting to a p2pool backend.

[-ck]

Don't forget that even if they end up being just one node in any p2pool network - be it the existing one or a new one, they will be far more than 51% of that network meaning it's as good as a centralised pool.

Like I said, they're paying homage to the idea of decentralisation because clients are demanding it so it makes for good marketing on their part. While them contributing to p2pool is most welcome, I see no sign of any of the alleged decentralisation.

[norgan]

Don't forget that even if they end up being just one node in any p2pool network - be it the existing one or a new one, they will be far more than 51% of that network meaning it's as good as a centralised pool.

Like I said, they're paying homage to the idea of decentralisation because clients are demanding it so it makes for good marketing on their part. While them contributing to p2pool is most welcome, I see no sign of any of the alleged decentralisation.

I think they intend on open sourcing the code so that existing nodes can then provide a "team" hashing effort while giving the node operator a way of paying out miners in that "team".
Problem there of course is we then have a burden of trust back on the miners towards the node operator.
I think its a step in the right direction though and a compromise between single central pools and the P2pool model.
That way you wouldn't have all miners on one node, you could still have local nodes to reduce latency. You would of course loose the freedom of moving between nodes without loosing your shares unless somehow the shares can be communicated to another node, a bit hard when the sharechain is tracking shares for the nodes wallet and not the miner itself.

[OVRGRO]

quick little FYI support tickets for antpool seem to go nowhere... i tripled my donation rate because how impressed i am with the customer service    shit at one point i was averaging over 1T for 10 minutes with 1 s3+, the other was pulling about 800g lol 1.8t with 2 s3 

[dogie]

quick little FYI support tickets for antpool seem to go nowhere... i tripled my donation rate because how impressed i am with the customer service    shit at one point i was averaging over 1T for 10 minutes with 1 s3+, the other was pulling about 800g lol 1.8t with 2 s3 

Seems to work for me, or do you mean a different page?

https://www.bitmaintech.com/about.htm

[dogie]

Don't forget that even if they end up being just one node in any p2pool network - be it the existing one or a new one, they will be far more than 51% of that network meaning it's as good as a centralised pool.

Like I said, they're paying homage to the idea of decentralisation because clients are demanding it so it makes for good marketing on their part. While them contributing to p2pool is most welcome, I see no sign of any of the alleged decentralisation.

The idea isn't to simply be the largest part of P2P, the idea is that they're investing time and money into developing the code for a pool sitting on p2p so others can do the same. As more people join P2P, they can put more hash rate onto P2P as appropriate.

I'm not sure exactly what option you're giving them to reach decentralisation otherwise. You want everyone on P2P -> but not if they're too big -> want them to do what?

[dogie]

Well, the node's been up for less than half an hour.  So much for their "extensive testing" claims.  Standard p2pool default front end.  Only thing different is they've exposed port 3333 instead of the typical 9332.

The non P2P pool is what was tested for months and months.

[BITMAIN]

Don't forget that even if they end up being just one node in any p2pool network - be it the existing one or a new one, they will be far more than 51% of that network meaning it's as good as a centralised pool.

Like I said, they're paying homage to the idea of decentralisation because clients are demanding it so it makes for good marketing on their part. While them contributing to p2pool is most welcome, I see no sign of any of the alleged decentralisation.

Hi ckolivas,

Yes p2p.antpool.com is only a node of p2pool network, but it brings more hashrate to the p2pool network. As mentioned in the announcement, we think we need to add hashrate to p2pool network first, to make the output more stable, then lots of p2pool nodes will show up. Finally, there will be a denctralized p2pool network.

[Solemn]

Well, I pointed a bit of hash power at AntPool a little over two days ago just to try it out, (Bitminter is my home) because I'm a "learn by doing" type of guy and really like the AntMiners I've got thus far.  Unfortunately, it's looking to be a very short run experiment.  So far, besides the incredibly inaccurate worker speed graph my biggest problem is that:

I HAVEN'T BEEN PAID!

I've been through two "automatic" payment cycles with balances above 0.01 with no joy. And now the settings pages doesn't even show the BTC address that I set up.  I have sent an email to info@bitmaintech as well as a support ticket (which appeared to disappear as soon as I selected "create, and is not listed on the left hand menu under "my tickets").  

I don't post much on here for a variety of reasons but this debacle has prompted me to reachout in a last ditch effort to get a bit of assistance from....well anyone related to bitmaintech...

Did I mention that there is no manual method for me to get my earnings, which I admit, wouldn't be an issue if the "automatic" payment system was working for me!  But it isn't...

-Edit:  14/11/14 - Just got an email from Bitmaintech asking when I started mining and repeating when the automatic payouts are supposed to happen. I've been mining since the 12th. and had enough to meet the min on the first day.  I can forgive the payout not happening on the first day...but it should have happened on the 14th!

[loshia]

Don't forget that even if they end up being just one node in any p2pool network - be it the existing one or a new one, they will be far more than 51% of that network meaning it's as good as a centralised pool.

Like I said, they're paying homage to the idea of decentralisation because clients are demanding it so it makes for good marketing on their part. While them contributing to p2pool is most welcome, I see no sign of any of the alleged decentralisation.

Hi ckolivas,

Yes p2p.antpool.com is only a node of p2pool network, but it brings more hashrate to the p2pool network. As mentioned in the announcement, we think we need to add hashrate to p2pool network first, to make the output more stable, then lots of p2pool nodes will show up. Finally, there will be a denctralized p2pool network.

Can i ask how much are you paying for your dedicated servers all over the world? And how many qualified people are involved in this project?  My rough estimate is that you are spending 3-5K USD monthly for your pool project.
The big question is:

How do you cover your costs having in mind it is 0% fee pool?
Thank you
PS: we all know that the Decentralization idea is very good but it brings you 0 USD in return!

[dogie]

How do you cover your costs having in mind it is 0% fee pool?

Transaction fees will cover some costs, but read the OP as to 'why' generally.

[dogie]

I HAVEN'T BEEN PAID!

And now the settings pages doesn't even show the BTC address that I set up.

That would be why, how can they pay you if there isn't an address on your account. Did you verify the address via the link sent by email?

[loshia]

How do you cover your costs having in mind it is 0% fee pool?

Transaction fees will cover some costs, but read the OP as to 'why' generally.

Come on dogie,
TX Fees
The fees can not pay even electricity for the servers alone. This is just a simple math. I do not have free time to loose to do the math/ but please feel free to prove me wrong

--------------------------
but read the OP as to 'why' generally.
-----------------------

I do not need to. I will leave this task to other folks