What's wrong with this picture? - page 2.

eleuthria

legendary

Activity: 1750

Merit: 1007

Quote from: organofcorti on June 14, 2012, 10:58:05 PM

I don't think DeepBit or Eligius are the problem - the problem is people want to use the network for free.

If that attitude doesn't change at some point and if paid transactions aren't given highest priority, then when block rewards cease the network hashrate will probably fall significantly and network security will suffer.

The problem is the re-definition of "Free". My understanding is that Tycho is looking at changing his pool's settings to include transactions that were being ignored. As of right now, Deepbit only allows a small number of "free" transactions, BUT Deepbit considers transactions using 0.0005 BTC/KB to be "free" currently, thus limiting how many it will put into a block. At this stage of the network, that seems like a problem since the DEFAULT value is 0.0005 BTC/KB.

There are other problems with redefining what a "free" transaction is at the current stage of Bitcoin development:
1) There is no easy way to add a fee to a transaction with bitcoind or bitcoin-qt. Basically bitcoin does it in the background, using the fee per KB you specified in configuration.
2) There is no easy way to know what fee your transaction will include (if any) until you try to send it using bitcoin-qt. Using bitcoind will send it and just tack on the fee without telling you in advance.

organofcorti

donator

Activity: 2058

Merit: 1007

Poor impulse control.

I don't think DeepBit or Eligius are the problem - the problem is people want to use the network for free.

If that attitude doesn't change at some point and if paid transactions aren't given highest priority, then when block rewards cease the network hashrate will probably fall significantly and network security will suffer.

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

I means that the pools: Eclipse and Ozcoin are putting through lots of transactions.
Other pools are not doing this.

Clearly some of the ones above with IP addresses are ignoring lots of transactions.

Deepbit and Eligius do ignore transactions that they consider to be "SPAM"
So they consider some of your transactions as "SPAM" coz you don't pay them enough
(or aren't in their 'good' list)

All the more reason to not use either of Deepbit or Eligius

hmongotaku

full member

Activity: 168

Merit: 100

Quote from: dlasher on June 13, 2012, 07:33:42 PM

Thank you EclipseMC and OzCoin...

what does it mean?

dlasher

sr. member

Activity: 467

Merit: 250

Thank you EclipseMC and OzCoin...

eleuthria

legendary

Activity: 1750

Merit: 1007

Quote from: DeepBit on June 08, 2012, 10:13:16 AM

Someone asked for my comments, so...

1. Block 183478 was found in just 12 seconds, so it was possibly a result from one of the very first getworks from this round. Those getworks contain only the coinbase TX because we had no time to build a complete block.
Block 183482 by other pool was found in 32 minutes, so it has lots of time to collect new TXes and include free or low-fee TXes not mined by Deepbit earlier.

2. Blockchaininfo's information about Deepbit's blocks is wrong sometimes. May be it's because other blocks are frequently received by them from our relaying nodes. Use the Blockorigin site, it takes information directly from Deepbit.
Looks like this time the block is really mine, and that's just our luck. Happens to anyone.

3. Usually Deepbit was including more free TXes than anyone else, but this situation changed recently. I heard from someone that those "1dice" TXes take up to 50% of the total numbers, and they are considered "free" by our fee policy, so we are limiting them in our blocks. There are lots of "1dice" TXes in that 747-TX block.
If this will become a problem, I'll consider rising my free TX limits.

Actually, 1dice TXes actually DO have fees attached, following the standard bitcoind fee assessment of 0.0005 per 1 KB (unless the coins are aged / many coins being sent, but that is not the case with most of dice transactions). As a matter of fact, I'm having trouble finding a single one in that 700 tx block that did not have a fee.

DeepBit

donator

Activity: 532

Merit: 501

We have cookies

Someone asked for my comments, so...

1. Block 183478 was found in just 12 seconds, so it was possibly a result from one of the very first getworks from this round. Those getworks contain only the coinbase TX because we had no time to build a complete block.
Block 183482 by other pool was found in 32 minutes, so it has lots of time to collect new TXes and include free or low-fee TXes not mined by Deepbit earlier.

2. Blockchaininfo's information about Deepbit's blocks is wrong sometimes. May be it's because other blocks are frequently received by them from our relaying nodes. Use the Blockorigin site, it takes information directly from Deepbit.
Looks like this time the block is really mine, and that's just our luck. Happens to anyone.

3. Usually Deepbit was including more free TXes than anyone else, but this situation changed recently. I heard from someone that those "1dice" TXes take up to 50% of the total numbers, and they are considered "free" by our fee policy, so we are limiting them in our blocks. There are lots of "1dice" TXes in that 747-TX block.
If this will become a problem, I'll consider rising my free TX limits.

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

Yeah if only BTC was like a credit card ... immediate commit and transaction fees.

Oh hang on, no, people want everyone to pay the transaction fees - but BTC doesn't have immediate commit for all transactions like a credit card ...

As for the picture at the top ... it's wrong - get over it.
blockchain.info is quite unreliable in determining the source of a block.
I can provide code to get 50% of blocks with 100% certainty, but the other 50% ... well ...

Assuming a block is from deepbit based on the IP address is clearly not reliable.

No doubt you saw the number at the bottom of that page about how many nodes they are connected to?
Currently: Status: Ok (260 Nodes Connected)
So yeah that's really gonna guarantee that the IP address is correct ... NOT.
Also a bitcoind could connect to deepbit and send its block to them first - and in that case the IP address method is clearly wrong

dlasher

sr. member

Activity: 467

Merit: 250

Quote from: Gladamas on June 08, 2012, 01:42:46 AM

I've found that even with the recommended fee of 0.0005 BTC for a transaction with few inputs, Deepbit still doesn't accept the transaction (I have to wait for other blocks.) If I do a fee of like 0.005 then it accepts it.

Quote from: TangibleCryptography on June 08, 2012, 01:37:55 AM

.....SatoshiDice generates a large # of tx and most (all?) have a tx fee so the % of free tx may have declined significantly....

Together, if true, this may be the problem we're presently seeing.

If SatoshiDice is generating scores of 0.0005 tx fee transactions (or less/free), and Deepbit isn't accepting anything below 0.005 (10 times higher) then we have a problem.

Gladamas

sr. member

Activity: 294

Merit: 250

Bitcoin today is what the internet was in 1998.

Quote from: Fuzzy on June 08, 2012, 01:15:48 AM

On one hand, they are under no obligation to calculate the free transactions, but on the other, it's discerning to see the largest pool taking this stance Undecided

I've found that even with the recommended fee of 0.0005 BTC for a transaction with few inputs, Deepbit still doesn't accept the transaction (I have to wait for other blocks.) If I do a fee of like 0.005 then it accepts it.

TangibleCryptography

sr. member

Activity: 476

Merit: 250

Tangible Cryptography LLC

Quote from: dlasher on June 08, 2012, 01:16:48 AM

Quote from: TangibleCryptography on June 08, 2012, 01:07:10 AM

Why? 95%+ of tx include no fee.

Not that I doubt you, but can you cite a source / graph / stats?

Sorry I can't just personal knowledge based on the % of tx I reject from the memory pool. My bitcoind is running patched version which has mintxfee option which I have set to 0.005 BTC. I guess it would have been useful to record that information. Maybe something for the future.

Then again that stats may be out of data. SatoshiDice generates a large # of tx and most (all?) have a tx fee so the % of free tx may have declined significantly.

I guess one way would be to use the blockchain.info API and parse the last 2016 blocks.

dlasher

sr. member

Activity: 467

Merit: 250

Quote from: TangibleCryptography on June 08, 2012, 01:07:10 AM

Why? 95%+ of tx include no fee.

Not that I doubt you, but can you cite a source / graph / stats?

Fuzzy

hero member

Activity: 560

Merit: 500

On one hand, they are under no obligation to calculate the free transactions, but on the other, it's discerning to see the largest pool taking this stance Undecided

TangibleCryptography

sr. member

Activity: 476

Merit: 250

Tangible Cryptography LLC

Quote from: dlasher on June 08, 2012, 12:49:19 AM

It does seem odd, as far as the network is presently behind, (3,111 unconfirmed transactions at the time of this message, 2322 low priority) that a block goes through with little to no transactions in it.

http://bitcoincharts.com/bitcoin/txlist/

Why? 95%+ of tx include no fee. Now if there are 3000+ paying tx which miss multiple blocks that is something to be concerned about. Personally I think it is awesome. We are starting to see differentiation in pay for service. If you want a fast confirm you are going to have to pay. If you don't care and just want it in a block eventually you might get away with a free tx.

dlasher

sr. member

Activity: 467

Merit: 250

It does seem odd, as far as the network is presently behind, (3,111 unconfirmed transactions at the time of this message, 2322 low priority) that a block goes through with little to no transactions in it.

http://bitcoincharts.com/bitcoin/txlist/

eleuthria

legendary

Activity: 1750

Merit: 1007

Quote from: Graet on June 07, 2012, 07:17:13 PM

Quote from: zerokwel on June 07, 2012, 07:11:20 PM

I don't think its about the blocks found and who finds them. (but then again I could be wrong)

But the lack of transactions within the blocks. Once again I could be wrong and tbh I don't know how pools work (I just mine). Is there a way pools can be lazy and not submit tx work.

I'm sure someone will be along soon with a LOT more brains than me

pools make money from txn fees, be silly not to include them eh?
Some pools refuse free txn - deepbit is not one

also blockchain.info is notoriously inaccurate with deepbit blocks - they may or may not even be deepbits,

look at
http://blockorigin.pfoe.be/top.php
for accurate block stats

Blockorigin is over 1k blocks behind Sad

. I missed their charts, it was nice to see a longer window (2016 blocks).

I have noticed the same thing lately though [I've been goofing around with SatoshiDice]. Blockchain.info will be showing 1000-1500 unconfirmed txes once in a while, and a Deepbit (it may be a false positive, but I doubt it is wrong EVERY time) block might only have 100-200 tx. A few minutes later, a block from another pool hits with 600-900. It's not a big problem right now, but it could become one later. If a pool is adding a large portion of the overall hash rate, but simply ignoring/not including a large number of transactions in each block, it could be a problem when we're seeing a higher transaction volume.

Graet

vip

Activity: 980

Merit: 1001

Quote from: zerokwel on June 07, 2012, 07:11:20 PM

I don't think its about the blocks found and who finds them. (but then again I could be wrong)

But the lack of transactions within the blocks. Once again I could be wrong and tbh I don't know how pools work (I just mine). Is there a way pools can be lazy and not submit tx work.

I'm sure someone will be along soon with a LOT more brains than me

pools make money from txn fees, be silly not to include them eh?
Some pools refuse free txn - deepbit is not one

also blockchain.info is notoriously inaccurate with deepbit blocks - they may or may not even be deepbits,

look at
http://blockorigin.pfoe.be/top.php
for accurate block stats

Snapman

sr. member

Activity: 291

Merit: 250

BTCRadio Owner

Is it possible for pool operators to only accept transactions with a txtfee greater than or equal to what they set?

zerokwel

sr. member

Activity: 464

Merit: 250

I don't think its about the blocks found and who finds them. (but then again I could be wrong)

But the lack of transactions within the blocks. Once again I could be wrong and tbh I don't know how pools work (I just mine). Is there a way pools can be lazy and not submit tx work.

I'm sure someone will be along soon with a LOT more brains than me

Tittiez

hero member

Activity: 686

Merit: 500

Nothing is wrong. All I see is a little variance.

We have these threads too often Tongue

Topic: What's wrong with this picture? - page 2. (Read 3764 times)