Topic: In very simple terms what's the hoopla about block size limit change (Read 619 times)

At this point I think we are stretching the analogy a bit too far to be useful for increasing understanding.  The analogy I stated was specifically to address the concept of an optimal supply for a given demand, as pointed out here:


Using an island population to represent a fixed demand, and a single farmer to represent an adjustable bitcoin network transaction supply. The farmer represents all the mining of the entire network, not just a single miner (or perhaps represents only the one farmer who sucessfully sells his crop that season? Whatever.), in the same way that the apple "profits" represent the total security of the entire network and not just a single miner's profits.

Beyond that there are too many nuances and details to make an analogy useful.  It quickly becomes just as complicated as stating the issue directly in terms of blocks, transactions, miners, bandwidth, electricity, hash power, protocols, nodes, fees, centralization, and adoption rate.

Nice analogy DannyHamilton. Let me expand it a bit to match the current debate.

In this fantasy world, there is a government that says that anyone is allowed to grow apples, but only 50 per week. If they grow any more than 50 per week they get kicked off the island and aren't allowed to sell any.

In addition the government says that all the farmers are only allowed to sell their apples on Sunday at the same market. The really weird thing is that the first farmer to make a sale the only one allowed to do business that day. The farmer is allowed to choose how many apples he sells. He might just sell 1 apple or he could sell all 50. All the other farmers go home empty handed.

The debate is to change the block size is basically saying all the farmers get to grow more than 50 apples per week. All the other rules stay the same.



My personal feeling is that right now, 50 apples are enough to feed the island, but eventually there will be thousands of people and 50 apples just won't be enough no matter how much you are will to pay per apple.

It just dawned on on me that there are two resources here that are in play.

Those with better internet bandwidth (lower bandwidth costs) and worse electricity (higher electric costs) would favor increased blocksize since they have the bandwidth resources needed.

Those with lower internet bandwidth (higher bandwidth costs) and better electricity (lower electric costs) would favor decreased blocksize since they have the electricity resources needed.

Lower electricity costs can drive the miners with higher electricity costs out of business by operating more hashes per electric dollar spent.

Lower bandwidth costs can drive the miners with higher bandwidth costs out of business by operating more hashes per bandwidth dollar spent.

The same bandwidth fear that people have about SUPERMINERS setting up shop with ULTRA-HIGH bandwidth and forcing centralization applies equally to SUPERMINERS setting up shop with ULTRA-CHEAP electricity and forcing centralization.

Certainly.

If it wasn't clear in my analogy, the "profits" of the apple tree owner represents the "security" of the bitcoin network (not the profits of the miners).

If I limit the transactions too much, I lose too many users, there aren't enough total fees per block and miners shut off their systems leading to a loss of security.
If I allow too many transactions, I expend too much resources on all the transactions that don't bring fees, miners shut off their systems leading to a loss of security.

The system won't work if I only allow 1 transaction per block.  The system won't work if I allow an infinite number of transactions per block.  Everyone has their own idea about what the "right" number of transactions per block is, with their own fears about how much resources are required to become fatal and how many transactions are needed to avoid being fatal.

Obviously as the population of the island grows (or the user base of bitcoin grows), the number of apples (transactions) that is optimum increases.

Anybody who thinks this is a simple problem with a simple solution is not paying attention.

If transactions don't get confirmed in a timely manner it doesn't matter how high the tx fees are. Bitcoin would die a horrible death.

Thanks for the wrapup, DannyHamilton.

Nice analogy.

But mining has one flaw there. As everyone can mine by simply buying hardware and investing electricity long term mining will always develop in the direction of a zero sum game. No matter how much fees are paid.

Therefore fees (long term) don't actually affect the profits of miners (very much) but only how much money is spend to secure the network.

Or in a analogy:

Just imagine in a community some guy shows up, offering to give 10,000 USD away every day. He will split this amount through everyone in this community that burns wood, depending on the amount of wood one burns.

What will happen? Everyone is going to buy more and more wood each day to burn it and get the biggest share of the 10,000 USD until closely to 10,000 USD worth of wood are burned each day.

Yes, as an analogy:

I own the only apple tree on an island.
My tree grows 100 apples in a season.
There are 50 residents on the island.

How many apples should I destroy to gain the highest profit?

If I destroy all but 1 apple, I might get a single bidder to pay $1000 for the apple while selling all 100 apples might only get me $0.10 per apple ($10) if nobody wants more than 2 apples and most people are willing to skip apples at a price higher than $0.10.

On the other hand, if I destroy all but 1 apple, I might find that the highest bidder on the island is not willing to pay more than $5 for the apple, while destroying half of the apples would mean I could find 25 people all willing to pay $0.25 each for 2 apples ($12.50)

Clearly there is an optimum number of apples that will bring an optimum price to maximize my profits.  (Note if I own an orchard on this island and produce thousands of apples per resident, most of the apples will rot and bring no revenue at all).

Limit the apples (transactions) too much and I might lose too many customers (users) and the system collapses.
Allow too many apples (transactions) and there is no incentive for anyone to cover my costs.  I'll probably find many apples (transactions) bringing no revenue (fees) at all.

Well said DannyHamilton. There is one additional point of the effect on miners.

If the blocksize gets bigger, it is expected that the fee per transaction will drop. There is a dispute if the net result will be an increase or decrease in reward for the miner. Essentially (Big Block * Little reward per transaction) = (Little Block * Big reward per transaction).

Satoshi himself wrote about a phased in solution for this issue:
https://bitcointalksearch.org/topic/m.15366

There is currently a hard limit on the size of each new block in the blockchain.  No single block can be more than 1MB in size.

As the total number of transactions occurring across the entire bitcoin network per minute continues to increase with the growing popularity of bitcoin, we are quickly going to reach a point where most of the time there are more transactions than will fit into a single block.  This means there will be a growing backlog of transactions waiting to be confirmed.

The current blocksize won't really support much more than 7 transactions per second.  So only the 7 people willing to pay the highest fee each second will get their transactions confirmed.  To encourage miners to confirm a transaction in their next block, there will essentially be a bidding war on transaction fees.  Some people are concerned about how high those fees might get as bitcoin continues to gain popularity.

Because of this there are some bitcoin developers who have suggested increasing the maximum blocksize beyond 1MB.  There are a variety of suggestions on how to do this including a suggestion for a new larger hard limit (meaning we might need this discussion again in the future), and a suggestion for a limit that would adjust automatically over time (risking an uncontrolled or unexpectedly large blocksize eventually depending on the criteria used for the adjustment).

If the blocksize gets too large, it will become impossible for those who are running on slower internet connections to participate.  It will most significantly effect miners who can't start mining the next block until they've received the most recently mined one.  This means that miners with very fast internet connections can feasibly push for extremely large blocks and force some of their competition to quit mining.  Repeat this cycle enough times with big enough blocks and eventually you lose a significant portion of the decentralization that provides bitcoin some of its security.  Make the block size big enough and it can begin to interfere with the ability of individuals to run a full node such as Bitcoin-Qt.

So, there is a tug of war between those who want to increase the number of transactions bitcoin can handle and keep transaction fees low, and those who want to ensure that we don't lose decentralization and full-node participation.

The change to blocksize requires a "hard fork" in the protocol.  This means that for it to work smoothly, it would require nearly 100% participation of all users (not just miners, but all users of all wallets).  If some percent of the users refuse to update to software that supports a larger blocksize, then the blockchain would split, and there would be two competing currencies both trying to be the official "Bitcoin".  To access the coins on both systems you'd have to run two separate wallets (one from each system).  Any coins that someone owns prior to the split could be spent twice (once on each of the splits).  Any coins mined (or received) on either of the split blockchains could only be spent on the one where it was mined (or received).

I've attempted to keep this explanation as unbiased as I could.  If I've failed in that effort and it appears that this description favors one side over the other, let me know and I'll attempt to re-phrase it.

Does this explain it well enough?

tl,dr;
Bigger block = more transactions, less fees, risk of decentralization (reducing security), risk of decreased full node participation (reducing security)

A few point and pros / cons would be very helpful for understanding this issue at hand.