Why a fee market and maybe even fees will never need to exist

[FUBAR-BDHR]

The whole blocksize debate with the limits and fee market has so many discussions going on about how fees will eventually need to replace the coinbase mining subsidy. There seems to be widespread agreement that fees will need to replace the subsidy at some point. I feel that this theory is incorrect and here is why.

If fees are kept low and bitcoin is allowed to scale on chain it will become more appealing for companies to adopt for every day use. More and more businesses will move toward accepting and using bitcoin. Some of those businesses will want and need to run their own nodes. Yes bigger blocks may cause some people with poor internet and insufficient funds to run a node to shut their nodes down. In the long run business nodes, academic nodes, and even government nodes will more than make up for the home nodes lost.

As the mining subsidy goes down it is foreseeable that if fees do not increase large mining operations will need to shutdown. I see this as a good thing. Businesses that use and rely on bitcoin (probably the ones running full nodes)can be made aware of this. In order to keep fees low and even free for themselves and their customers they can fire up their own small mining operations in their own data centers. Yes I said free. They can include their own transactions into blocks the mine for free and partner with other miners to do the same. Mining will again become decentralized. Maybe these small business data centers are what Satoshi foresaw.

Why would companies pay for these mining operations? They pay huge amounts in fees for bank and credit card transactions. They could easily save on those fees with bitcoin and redirect that cash to mining and nodes. If not greedy they could even offer small discounts to bitcoin customers and still be ahead.

Imagine if every fortune 500 company ran at least one node and either ran a node for themselves or paid a data center for hash rate to keep their fees at or near 0. Right now according to my irc bot the network hash rate is about 2368ph. That's 4.736ph per company. Now more than just those companies will want this benefit so cut that number way down. People with cheap and free electric will also cut that down as long as they can pick up or continue to run old miners.

Afraid of the size of blocks needed for all those transaction? Think the node count will be a problem with blocks ? There were 26355 universities worldwide in 2016[1]. Add that to businesses that need nodes and even a small fraction of the libraries in the world and you are over 50000 nodes easily. That enough for decentralization?

Another reason fees never need to increase or can even go down. As long as the price of bitcoin doubles every four years there is no need for mining fees. At the first halving the price was $12.25 and at the second $652.14[2]. Do you see $1304.28 by 2020 and 2608.54 by 2024? That's all it takes for fees to be minimal or 0 and mining rewards to have the same value.

So I give to you my version of the future where nodes are plentiful, mining returns to decentralized, and fees are not needed. No fee market needs to develop. We just bigger blocks to increase transactions per second to allow for that adoption to occur.



[1] http://www.webometrics.info/en/node/54
[2] https://99bitcoins.com/price-chart-history/

[Decoded]

What you're proposing is basically an indirect fee. Instead of paying fees to the miners, you want to force the bigger operations to shut down (Which are kinda holding the Bitcoin price up), so smaller companies can run their own mining operation to prioritise smaller transactions. I think it's stupid.

Let's look at it from a different perspective. Let's say in 2140, each company listed on every worldwide stock exchange runs a node. An estimate is double the number we have now, so almost 100000 companies. If you want to make a transaction, you are most definitely not sending 100000 people money for a chance that ONE of them will mine your transaction for you. And you're definitely not paying just ONE of them, because what are the odds of that one miner mining the next block? It's just not plausible. You're paying much more.

My point is - Each person running a mining node and prioritising their own transactions will not work. You paying other people to put your transaction in next will not work, because you don't know who will mine the next transaction, and you are not paying every single miner.

[Decoded]

You're saying that it the price of Bitcoin doubles every halving, then the miners will stay making the same amount. Can you see how horribly wrong you are?

Last halving, the price was $600. Let's say there was a 2 btc transaction reward too. This is very generous, it almost never hit 2btc. So each block was 27*600 USD.

Right now, let's say the price rose to $1200, and as you say, the fees are kept the same. 14.5*1200 USD. Hey, it seems alright.

But then the next year it keeps going. 8.25*2400.

5.125*4800.

3.5625*9600.

2.78125*19200.

Going this far in, it's obvious that this will not be sustainable. A normal fee of 0.0003 btc and at $19200 per Bitcoin, it will cost almost $6 to send a transaction. With the more inputs and outputs, it will only scale higher and higher.

The fees in Bitcoin may stay the same, but it will only grow more and more in fiat. This ain't even counting inflation.

[FUBAR-BDHR]

What I am saying is that once adoption occurs that companies will mine and run node to support their own best interest. Any fee will eventually become unnecessary. Well maybe a min fee will be necessary for normal user transactions that do not involve a company that mines to prevent spam attacks, Corporations can send 0 fee transactions for themselves and their customer just like mining pools do. Companies will either rent hash kind of like cloud mining or directly own their miners. They will form networks with partners to ensure shorter wait times for their 0 fee transactions and prioritize those transactions. It all relies on adoption by big companies which cannot happen with small blocks. It may take years or decades for the slow migration to happen but I see it happening in the future.

[Decoded]

What I am saying is that once adoption occurs that companies will mine and run node to support their own best interest. Any fee will eventually become unnecessary. Well maybe a min fee will be necessary for normal user transactions that do not involve a company that mines to prevent spam attacks, Corporations can send 0 fee transactions for themselves and their customer just like mining pools do. Companies will either rent hash kind of like cloud mining or directly own their miners. They will form networks with partners to ensure shorter wait times for their 0 fee transactions and prioritize those transactions. It all relies on adoption by big companies which cannot happen with small blocks. It may take years or decades for the slow migration to happen but I see it happening in the future.

The thing is, that this is infeasible. Compared to the current ease of paying small fees to get your transaction in, you have to run your own node, and rent hashpower? Let's say to get a decent chance of getting into the next block, we need 30% of the network hashpower. This means it would take on average a little more than half an hour to get your 0 fee transaction confirmed, if you are the miner. At the moment, there is about 2.5 petahashes of hashing power, meaning that you need around 0.8 petahashes.

You can fit about 3500 transactions max into a 1 MB block. Let's be extremely optimistic, and say each company will only use one transaction per block. So If you have a clump of 350 companies, which isn't realistic, you would need a datacenter purely run on Antminer S9s (The best on the market right now, the most value per hash) which run on 14 TH/s, you will need just under 60 of them. One of these cost about 2100 USD, so in total, just the miners cost $360 per company. Add the electricity costs, housing costs, upkeep labor costs and equipment costs, it will be $500 initially, then more than $100 a year. Doesn't sound very nice.