[bitcoin-dev] High fees / centralization

[Tom Harding via bitcoin-dev]

Raystonn,

Your logic is very hard to dispute. An important special case is small
miners.

Small miners use pools exactly because they want smaller, more frequent
payments.

Rising fees force them to take payments less frequently, and will only tend
to make more of them give up.

With fees rising superlinearly, this centralizing effect is much stronger
than the oft-cited worry of small miners joining large pools to decrease
orphan rates.


On Mar 29, 2017 15:01, "Raystonn . via bitcoin-dev" <
bitcoin-***@lists.linuxfoundation.org> wrote:

Low node costs are a good goal for nodes that handle transactions the node
operator can afford. Nobody is going to run a node for a network they do
not use for their own transactions. If transactions have fees that
prohibit use for most economic activity, that means node count will drop
until nodes are generally run by those who settle large amounts. That is
very centralizing.

Raystonn

On 29 Mar 2017 12:14 p.m., Jared Lee Richardson via bitcoin-dev <
bitcoin-***@lists.linuxfoundation.org> wrote:

In order for any blocksize increase to be agreed upon, more consensus is
needed. The proportion of users believing no blocksize increases are
needed is larger than the hardfork target core wants(95% consensus). The
proportion of users believing in microtransactions for all is also larger
than 5%, and both of those groups may be larger than 10% respectively. I
don't think either the Big-blocks faction nor the low-node-costs faction
have even a simple majority of support. Getting consensus is going to be a
big mess, but it is critical that it is done.

[David Vorick via bitcoin-dev]

On Mar 30, 2017 12:04 PM, "Tom Harding via bitcoin-dev" <
bitcoin-***@lists.linuxfoundation.org> wrote:

Raystonn,

Your logic is very hard to dispute. An important special case is small
miners.

Small miners use pools exactly because they want smaller, more frequent
payments.

Rising fees force them to take payments less frequently, and will only tend
to make more of them give up.

With fees rising superlinearly, this centralizing effect is much stronger
than the oft-cited worry of small miners joining large pools to decrease
orphan rates.


Miners get paid on average once every ten minutes. The size of fees and the
number of fee transactions does not change the payout rate.

Further, we are very far from the point (in my appraisal) where fees are
high enough to block home users from using the network.

Bitcoin has many high-value use cases such as savings. We should not throw
away the core innovation of monetary sovereignty in pursuit of supporting
0.1% of the world's daily transactions.

[Jared Lee Richardson via bitcoin-dev]

Post by David Vorick via bitcoin-dev
Further, we are very far from the point (in my appraisal) where fees are

high enough to block home users from using the network.

This depends entirely on the usecase entirely. Most likely even without a
blocksize increase, home purchases will be large enough to fit on the
blocksize in the forseeable future. Microtransactions(<$0.25) on the other
hand aren't viable no matter what we try to do - There's just too much data.

Most likely, transaction fees above $1 per tx will become unappealing for
many consumers, and above $10 is likely to be niche-level. It is hard to
say with any certainty, but average credit card fees give us some
indications to work with - $1.2 on a $30 transaction, though paid by the
business and not the consumer.

Without blocksize increases, fees higher than $1/tx are basically
inevitable, most likely before 2020. Running a node only costs $10/month
if that. If we were going to favor node operational costs that highly in
the weighting, we'd better have a pretty solid justification with
mathematical models or examples.

Post by David Vorick via bitcoin-dev
We should not throw away the core innovation of monetary sovereignty in

pursuit of supporting 0.1% of the world's daily transactions.

If we can easily have both, why not have both?

An altcoin with both will take Bitcoin's monetary sovereignty crown by
default. No crown, no usecases, no Bitcoin.



On Thu, Mar 30, 2017 at 9:14 AM, David Vorick via bitcoin-dev <

Post by David Vorick via bitcoin-dev
On Mar 30, 2017 12:04 PM, "Tom Harding via bitcoin-dev" <
Raystonn,
Your logic is very hard to dispute. An important special case is small
miners.
Small miners use pools exactly because they want smaller, more frequent
payments.
Rising fees force them to take payments less frequently, and will only
tend to make more of them give up.
With fees rising superlinearly, this centralizing effect is much stronger
than the oft-cited worry of small miners joining large pools to decrease
orphan rates.
Miners get paid on average once every ten minutes. The size of fees and
the number of fee transactions does not change the payout rate.
Further, we are very far from the point (in my appraisal) where fees are
high enough to block home users from using the network.
Bitcoin has many high-value use cases such as savings. We should not throw
away the core innovation of monetary sovereignty in pursuit of supporting
0.1% of the world's daily transactions.
_______________________________________________
bitcoin-dev mailing list
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev

[Vladimir Zaytsev via bitcoin-dev]

There must be a way to organize “branches” of smaller activity to join main tree after they grow. Outsider a bit, I see going circles here, but not everything must be accepted in the chain. Good idea as it is, it’s just too early to record every sight
.

Further, we are very far from the point (in my appraisal) where fees are high enough to block home users from using the network.

This depends entirely on the usecase entirely. Most likely even without a blocksize increase, home purchases will be large enough to fit on the blocksize in the forseeable future. Microtransactions(<$0.25) on the other hand aren't viable no matter what we try to do - There's just too much data.
Most likely, transaction fees above $1 per tx will become unappealing for many consumers, and above $10 is likely to be niche-level. It is hard to say with any certainty, but average credit card fees give us some indications to work with - $1.2 on a $30 transaction, though paid by the business and not the consumer.
Without blocksize increases, fees higher than $1/tx are basically inevitable, most likely before 2020. Running a node only costs $10/month if that. If we were going to favor node operational costs that highly in the weighting, we'd better have a pretty solid justification with mathematical models or examples.

We should not throw away the core innovation of monetary sovereignty in pursuit of supporting 0.1% of the world's daily transactions.

If we can easily have both, why not have both?
An altcoin with both will take Bitcoin's monetary sovereignty crown by default. No crown, no usecases, no Bitcoin.

[Jared Lee Richardson via bitcoin-dev]

That would be blockchain sharding.

Would be amazing if someone could figure out how to do it trustlessly. So
far I'm not convinced it is possible to resolve the conflicts between the
shards and commit transactions between shards.

On Thu, Mar 30, 2017 at 6:39 PM, Vladimir Zaytsev <

Post by Vladimir Zaytsev via bitcoin-dev
There must be a way to organize “branches” of smaller activity to join
main tree after they grow. Outsider a bit, I see going circles here, but
not everything must be accepted in the chain. Good idea as it is, it’s just
too early to record every sight
.
On Mar 30, 2017, at 5:52 PM, Jared Lee Richardson via bitcoin-dev <

Post by David Vorick via bitcoin-dev
Further, we are very far from the point (in my appraisal) where fees

are high enough to block home users from using the network.
This depends entirely on the usecase entirely. Most likely even without a
blocksize increase, home purchases will be large enough to fit on the
blocksize in the forseeable future. Microtransactions(<$0.25) on the other
hand aren't viable no matter what we try to do - There's just too much data.
Most likely, transaction fees above $1 per tx will become unappealing for
many consumers, and above $10 is likely to be niche-level. It is hard to
say with any certainty, but average credit card fees give us some
indications to work with - $1.2 on a $30 transaction, though paid by the
business and not the consumer.
Without blocksize increases, fees higher than $1/tx are basically
inevitable, most likely before 2020. Running a node only costs $10/month
if that. If we were going to favor node operational costs that highly in
the weighting, we'd better have a pretty solid justification with
mathematical models or examples.

Post by David Vorick via bitcoin-dev
We should not throw away the core innovation of monetary sovereignty in

pursuit of supporting 0.1% of the world's daily transactions.
If we can easily have both, why not have both?
An altcoin with both will take Bitcoin's monetary sovereignty crown by
default. No crown, no usecases, no Bitcoin.

[Vladimir Zaytsev via bitcoin-dev]

Can there be a minimum amount to put up for mining ? I hope i’m not in violation with any ideology yet :)

Post by Jared Lee Richardson via bitcoin-dev
That would be blockchain sharding.
Would be amazing if someone could figure out how to do it trustlessly. So far I'm not convinced it is possible to resolve the conflicts between the shards and commit transactions between shards.
There must be a way to organize “branches” of smaller activity to join main tree after they grow. Outsider a bit, I see going circles here, but not everything must be accepted in the chain. Good idea as it is, it’s just too early to record every sight
.

Further, we are very far from the point (in my appraisal) where fees are high enough to block home users from using the network.

This depends entirely on the usecase entirely. Most likely even without a blocksize increase, home purchases will be large enough to fit on the blocksize in the forseeable future. Microtransactions(<$0.25) on the other hand aren't viable no matter what we try to do - There's just too much data.
Most likely, transaction fees above $1 per tx will become unappealing for many consumers, and above $10 is likely to be niche-level. It is hard to say with any certainty, but average credit card fees give us some indications to work with - $1.2 on a $30 transaction, though paid by the business and not the consumer.
Without blocksize increases, fees higher than $1/tx are basically inevitable, most likely before 2020. Running a node only costs $10/month if that. If we were going to favor node operational costs that highly in the weighting, we'd better have a pretty solid justification with mathematical models or examples.

We should not throw away the core innovation of monetary sovereignty in pursuit of supporting 0.1% of the world's daily transactions.

If we can easily have both, why not have both?
An altcoin with both will take Bitcoin's monetary sovereignty crown by default. No crown, no usecases, no Bitcoin.

[Staf Verhaegen via bitcoin-dev]

Jared Lee Richardson via bitcoin-dev schreef op do 30-03-2017 om 19:01

Post by Jared Lee Richardson via bitcoin-dev
That would be blockchain sharding.
Would be amazing if someone could figure out how to do it trustlessly.
So far I'm not convinced it is possible to resolve the conflicts
between the shards and commit transactions between shards.

I'm thinking more of a system where different nodes can agree to do part
of the transaction processing. In that way 20000 nodes could work like
5000 full nodes.

greets,
Staf.

[Tom Harding via bitcoin-dev]

Post by David Vorick via bitcoin-dev
On Mar 30, 2017 12:04 PM, "Tom Harding via bitcoin-dev"
Raystonn,
Your logic is very hard to dispute. An important special case is
small miners.
Small miners use pools exactly because they want smaller, more
frequent payments.
Rising fees force them to take payments less frequently, and will
only tend to make more of them give up.
With fees rising superlinearly, this centralizing effect is much
stronger than the oft-cited worry of small miners joining large
pools to decrease orphan rates.
Miners get paid on average once every ten minutes. The size of fees
and the number of fee transactions does not change the payout rate.
Further, we are very far from the point (in my appraisal) where fees
are high enough to block home users from using the network.
Bitcoin has many high-value use cases such as savings. We should not
throw away the core innovation of monetary sovereignty in pursuit of
supporting 0.1% of the world's daily transactions.

Owners of small mining rigs get paid by pools, generally using regular
transactions that pay regular fees (p2pool is an exception that pays
directly from coinbase). The point is the unintended consequences are
directly at odds with one of the justifications offered for small blocks
- miner centralization.

This is a special case. Raystonn's general point was that high fees
will lead to fewer economic actors overall, and therefore fewer full nodes.