Topic: Fundamental fee / security issue in bitcoins future? (Read 1481 times)

You can still do all those things. Just take the precautions mentioned in the wiki for the Race Attack and you'll have some measure of protection. You just won't have as much as if you took more proactive measures.

As far as paying fees goes, somebody is going to end up paying them to someone. Perhaps not very much per transaction, or for all of them, but one way or another miners are going to get paid. Will the fees be paid entirely by end users? Probably some of them will but I see a business case for merchants to either set up their own pools, or set up agreements with existing pools to process their customers' transactions with mining being funded out of band.

When bitcoin grows there's not going to be just one way of using it, just like there's not just one way of using the Internet.

This means that if you want to put a bitcoin point of sale terminal in your store, instead of do it yourself, trust no one, pay no fees to banks approach, suddenly you have to talk to a guy in Prague named Slush.

Currently, about a million bitcoins per year goes for security.  After 2140, and in practice much sooner, if you assume 1% transaction fees, to get the same level, you need 100 million in volume.  

This means the security will probably go down.  Parked bitcoins become freeloaders on the network.  The network lives or dies depending on the transaction volume.  Miners stop accepting free transactions.


There is also added complexity of bitcoin price.  If most of the bitcoins is parked, value of those bitcoins (as well as all others) is determined by a small part that transacted.  

Transaction fees will always push the bitcoin exchange price down, and if there are no other factors, bitcoin price will tend to zero.  By "no other factors" in this case I mean no speculative buying and selling, and no round trip transactions.  This is a silly case, because there is nobody buying.

If we add round trip transactions, the stable price is where the value added trough both kind of transactions is equal to the mining fees, and thus indirectly, security.  More value added, more security.

If we magically again remove round trip transactions and add speculators and savers, the stable state is where fees equals new speculative amount.  Bitcoins keep value as long as all saver keep buying a percentage of their value each year.  If they don't do it, bitcoin price again goes down.

Another problem with bitcoin as a store of value where the security is determined by a minority that pays transaction fees, is that the liquidity is provided by the round tripping part, and they are disconnected from the price.  So it means that hoarders save on the fees but do get hit by the volatility they themselves are creating.

The 51% attack theory is in all likelyhood now no longer possible unless it were organized by a mining pool or a massive botnet. Bitcoincharts lists the computing power behind the Bitcoin network at 703 PetaFLOPS right now. The most powerful (known) supercomputer in the world, the Department of Energy's Titan at Oak Ridge, runs at about 17.5. In fact, the 10 most powerful (known) supercomputers combined would amount to only about 30% of what BTC Guild can do by themselves.

All of your other concerns I think will eventually be mooted when the major payment networks realize they can leverage the technology better. These companies have experience building the UI, and despite their warts, are far more trustworthy in the public eye than an InstaWallet.

We've already gone from used $100 ATIs off of eBay to dedicated ASIC rigs that cost five-figures in the span of less than a year. By 2015, mining-at-home as a hobby will probably be dead, since all the hashpower will be centralized by existing payment processors that adopt bitcoin or VC funded startups.

Have you seen the complaints about fees being too large?

If people hoard, that means the price will go up. As the price goes up the fees gain in value. As the fees gain in value, more people will mine.

Also, you are talking about 2140. We have to first deal with the problem of transaction time between Earth and Mars.

That sounds ominous, but upon further examination I have no idea what it actually means.

That just means you are solving problems in a distributed system by more centralization.

In other words, bitcoin is a bad long term store of value.

As Bitcoin grows businesses will become involved in mining, either directly, or as part of a merchant pool, or by buying premium services from existing pools.

The biggest risk for zero confirmation payments is the race attack, but if I have confirmation from the top 3-5 mining pools that the transaction got into their memory pool before a double spend transaction then I can be reasonably certain it will go through.

Large pools could earn extra money by offering a service that either allows subscribers to instantly check the status of a particular transaction vis-à-vis the memory pool and/or guarantees that certain transactions will be included in the next block.

Ultimately I expect mining to cease to exist as a standalone industry and to become something that businesses do themselves in order to make sure their transactions get processed. Just like how every business of significant size has an accounting department, in the future every business will have a bitcoin mining department.

As Bitcoin 'grows' the incentive for businesses to add transaction fees to their model business will be rather important.

Would you as a consumer prefer to buy from a business that will take 2-4hrs confirming and selling you your purchase? Or would you prefer from a business that confirms your purchase (holiday/house/car/monkey/etc) within a matter of minutes (possibly faster)?

A business can not 'grab' part of those transaction fees, but it is most definitely in a business best interest to add them on as part of their charge (maybe not telling the customer, but just paying the transaction fee themselves).

It is quite possible that it would become part of a business model to openly show the transaction fee they put in, or they might offer (as someones mentioned in another thread). Especially with the size of the 'Bitcoin market' in the future.

Transactions fees are only destined to grow, exponentially.

As fees grow, miners grow, balancing it back out and increasing the combined network power, again, exponentially (but you know this already).

And as mentioned, it is in the best interests of the miners to spend their bitcoin.

If they save xxx bitcoin, then bitcoin loses value because of xxx missing becomes noticable and people start trading in another currency type, then Miner A's vast store of bitcoins suddenly starts dropping in value (all that time, money spent on computer resources, etc, wasted), which pulls the hoarder into spending his/her bitcoin.

If hoarder wants to actually keep their bitcoin until its value drops stupidly, then before this occurs (because we, the people, have noticed a huge amount missing and being stored) a majority agreement concurs to 'mint' new bitcoins, bringing energy back to the bitcoin for everyone else AND devalue'ing even further the hoarders stash.

These two contradict. If majority of coins are hoarded as a store of value, then the gain of a 51% attack will be minimum

Overall, 51% attack is not a big deal, you can only disturb the network for a while and make a couple of double spend, which might be observed anyway, for existing coin hoarders, 51% attack won't affect them

The incentive for mining in the future is a question, but it is too uncertain to discuss it now without a roughly idea how the future environment looks like

See here:

https://bitcointalksearch.org/topic/funding-of-network-security-with-infinite-block-sizes-157141

Am I correct in my calculations that only between 1 and 2% of what miners are earning is currently coming from fees? (roughly based on this chart: http://blockchain.info/charts/transaction-fees ) and the current (I think?) 3600-coins-per-day inflation rate.

It's a valid concern, but there's no way to tell if it's a fatal flaw or predict how it will play out over time. If bitcoin catches on for payments, there will be zillions of them going thru the network, and as any billion dollar credit card processor will tell you, all those small fees add up over time.

BTW, it's fees not fee's.

Also, there will be miners who just like to mine right now, for the sake of mining. Not necessarily for profit. As an individual, I might want to mine at the current difficulty if I could get a low cost ASIC to enable me to earn 1 BTC per day (right now, that would mean about 15 GH/s).

This is not a problem until maybe the next 10 years. By then, we'd see if this is still a problem.

You're assuming that miners are completely distinct from the people who want the network to be secure (users/merchants/exchanges/etc).

That is a bad assumption. Nothing stops a merchant who wants more network security from either subsidizing miners (maybe in exchange for a promise to prioritize transactions to them) or mining themselves.

This is already happening, not for reasons of security but for other reasons.

I'm seeing some fundamental issues in bitcoins future from my current understanding. I could certainly be wrong, but I'd love to talk about it and find out why I'm wrong if I am. I'll start out with my premises and go from there.

1 In the future, the majority of miners profits will be derived from fee's once bitcoin's reward depletes and all coins are issued.

Fairly self evident, no coins means miners need to make their money somewhere.

2 The amount of hashing power in the system directly correlates to the security of the system, as it determines the computational power required to create a 51% attack.

Again I believe this is self evident. If the system total hash is higher you need a greater amount of hashing power to perform a 51% attack.

3 The amount of hashing power in the system directly correlates to the money available to be made via transaction fee's.

This is a ratio compared to what is available on the market for H/$, not so much a fixed amount. For example if in the future super efficient ASICs are developed this could be something like 1 GH/$, the number itself is irrelevant. The key point is that miners will always stabilize around the cost of getting a finger in the pie against their reward for having a slice of pie. If it costs more to run then you get rewarded miners unplug. If it's cheaper to get in and it's profitable to get rewarded miners will do so.

4 The greater market cap in bitcoins, the more attractive it is for the target of a 51% attack, as there is more to gain

I think this is fairly self evident, being able to control a 100 billion dollar market is more advantageous than controlling a 1 billion dollar market, which is more beneficial than a 1 million dollar market, etc.


From these premise I draw the conclusion that the amount of hashing power in the system (premise #3) will directly correlate with the amount of transactional fee’s in the system (premise #1), and hence the security of the system (premise #2). I also believe there is a ratio between the market cap, and the transactional fee’s being generated (MC:T ratio). The issue becomes problematic if BTC is ever used to store wealth rather than make transactions as a currency. We see in this situation that the MC:T ratio would get larger and larger, with wealth being pored into a closed system, yet there not being enough transactions to generate the security required for the increased size of the market. According to premise #4 we have an increasing target, but not the security to go along with it.

So – What can we do about it? Honestly I can’t really think of much without destroying the bitcoin. You could increase the fee rate per transaction, but then you’d be destroying one of the key strengths of the bitcoin – cheap ways of transmitting money from person to person. It’s possible that the bitcoin could grow large enough that the transaction fee’s could be valuable enough to stop a 51% attack, even if there is an obscenely large MC:T ratio, but I wouldn’t like to count on that idea, as if something is incredibly profitable they will find a way to exploit it.

In my opinion if the majority of coins are horded instead of used as a currency (i.e. wealth storage) then the bitcoin could have long term issues, and possibly be subjected to attack.

TL;DR Currently bitcoin hash rate is tied to worth because you're rewarded in a fashion directly related to their worth. In the future hash rate is tied to transaction frequency, which is independent of the network worth and hence need for security.

Thoughts?