Questions about scalability and growing computing power

[Wouter Drucker]

If Bitcoin would become the n1 currency in the world, would this simply be too much to process? I read the wiki which gives me the feeling that it's not so bad, but what I get from the community, and from my understanding of what Bitcoin is, it seems scalability is a major weakness.

My other question is this. It seems we are going to see a network develop with a massive amount of computing power. We have both Moore's law and miner competition working at the same time. This way it becomes harder to attack the network. But it seems like a waste of computing power, computers are performing (senseless) puzzles that just become harder and harder. Wouldn't it make more sense if some of that growing computer power where to go to make the network scale (make it faster and allowing it to handle more traffic)?

[acoindr]

Well, in the first place I don't think Bitcoin will become the #1 currency in the world. I think that will revert back to some sort of gold standard. Bitcoin is more like a money cop keeping things (governments) honest, which is sorely needed and why, absent some unforeseen flaw, it will always have value.

I do think Bitcoin will have a global footprint. There has been a lot of discussion about how the network could accommodate a global user base. One thing to remember is not all Bitcoin transactions need to happen on the blockchain. I wrote a post to get thoughts and discussion going on off-chain transactions, which of course can scale quite easily. Additionally, core development discussion has been in the direction of allowing the core network to accommodate such volume too. This is certainly possible technically.

I'm not really worried about scalability because I think any rough spots will be smoothed out as things progress and evolve. For example, Bitcoin is not alone. There is already Litecoin, which has probably near the same number of holders as Bitcoin by now, and provides crypto transaction back up. So I believe there will be a combination of off-chain solutions, alt-coin options, and technology progress which ultimately smooths scalability out quite nicely. Remember going to the moon once seemed impossible (and certainly going to Mars would be) but as people focused on the problem things became easier and one day quite doable.

As for usage of network computing, mining performs the very necessary and valuable function of securing and auditing transactions, so that's not a waste. Additionally, there have been other ideas proposed to harness some of the network computing power and some miners use their rigs to heat their homes (seriously).

[Wouter Drucker]

Wouldn't you get an arms race resulting in ever increasing security, while at some point extra security would be redundant? If miners are now paid for adding security, wouldn't it make more sense to pay them for adding speed once security is at a high enough level?

[acoindr]

Wouldn't you get an arms race resulting in ever increasing security, while at some point extra security would be redundant? If miners are now paid for adding security, wouldn't it make more sense to pay them for adding speed once security is at a high enough level?

I think you have a misunderstanding of how Bitcoin works.

You can't add transaction speed. When you make a transaction on the core Bitcoin network it gets immediately broadcast and (if valid) relayed to the rest of the nodes participating in Bitcoin. This happens nearly instantly. So if you send me one of your bitcoins I'll get that in my wallet quickly whether I'm sitting right next to you or across the world. The problem is knowing if I can trust that the transaction will be seen as valid by the network forever into the future. To know that I need to wait for what's called "confirmations".

Confirmations happen when the transaction is included into a block by a miner on the network. It is widely held that after 6 confirmations (6 blocks that acknowledge the transaction) it can be regarded as permanently valid because to undo that requires immense computing power, and more so with further confirmations.

The time between Bitcoin blocks is automatically adjusted to remain at a target of 10 minutes to keep the number of coins coming into circulation relatively steady. So to have a fully confirmed Bitcoin transaction takes on average 1 hour on the core network. This is another reason I believe off-chain transactions, which can be confirmed as valid instantly, will play a big role in Bitcoin's future.

The security of the network (difficulty in attacking it) automatically rises or lowers depending how many miners participate. As Bitcoin adoption grows the number of mining participants, and therefore security, naturally rises.

[Hawker]

Doesn't the number of miners fall as ASIC type machines become the norm?

[acoindr]

Doesn't the number of miners fall as ASIC type machines become the norm?

Not really, although the composition of miners may look different.

ASICs are simply specialized equipment for mining. Bitcoin resembles gold in many ways and dependency on the free market is a big one.

There are all sorts of gold miners, from ones using simple pans to large industrial ones with expensive specialized equipment. If gold rose significantly in value and demand so to would the composition of people mining it, but there would probably be more overall.

The same would apply with Bitcoin. When bitcoins were worth $0.10  each you could mine them easily with a rinky dink desktop computer. As their value increased over $1.00 more expensive machines and rigs were employed like racks of GPUs. As bitcoin value went into double digits ASICS began being developed. Bitcoin value will likely continue to rise and the competition for them and makeup of people mining them will continue to change, but probably always be more overall.

[Hawker]

Fair point.

[Wouter Drucker]

But couldn't the verification process be sped up over time?

[Come-from-Beyond]

If Bitcoin would become the n1 currency in the world, would this simply be too much to process? I read the wiki which gives me the feeling that it's not so bad, but what I get from the community, and from my understanding of what Bitcoin is, it seems scalability is a major weakness.

Yes. Bitcoin has scalability issues, but developers work hard to solve them.

[acoindr]

But couldn't the verification process be sped up over time?

As I described above Bitcoin uses a target time of 10 minutes between blocks. The only way to speed up verification for using Bitcoin with the blockchain (as opposed to off-chain transactions) would be to make a protocol change reducing the time between blocks. There is no large priority to do this. For example, even if you cut the time in half to 5 minutes that's still 30 minutes for full 6 block confirmation. If you're using bitcoins to pay for a cup of coffee somewhere, you probably would rather use an off-chain transaction solution so it can be confirmed instantly.

Note that one of the features of Litecoin, however, is a reduced target time between blocks of about 2.5 minutes, so that means a full 6 block confirmation time of about 15 minutes.

[adamthefishman]

If you were buying something for 1million dollars, 6 confirmations would probably be required.

For a small transaction however, how hard is it to fake a bitcoin transfer?

Would anyone attempt it to steal a coffee?

After 6 confirmations, the transaction is 100% safe.

What about after 5, or even after 1?

[acoindr]

If you were buying something for 1million dollars, 6 confirmations would probably be required.

For a small transaction however, how hard is it to fake a bitcoin transfer?

Would anyone attempt it to steal a coffee?

After 6 confirmations, the transaction is 100% safe.

What about after 5, or even after 1?

Yes, you're right that there is a risk factor versus amount that can be taken into consideration. The Bitcoin network is so large now many people feel safe with less than 6 confirmations for some items. For example, BTC-e credits your account after only 3 confirmations. For a transaction only worth a few dollars 1 confirmation is probably plenty.

[Justin00]

I still can't really see the incentive for stores (not online stores, but actual stores you walk into) to accept BTC.
Also unless you mine BTC, why would you buy BTC for $$ and then use the BTC to buy something from a store ?

[acoindr]

I still can't really see the incentive for stores (not online stores, but actual stores you walk into) to accept BTC.
Also unless you mine BTC, why would you buy BTC for $$ and then use the BTC to buy something from a store ?

Stores (of all types) have incentive and desire to make sales. Regardless what you sell if a growing number of your potential customers ask if you accept a certain payment type, pretty soon you're likely to say yes.

As Bitcoin adoption grows, as it is obviously now, that means more likelihood of more places accepting it, as is happening now. As more people accept it, more people have it, meaning more people will want it accepted more places. This cycle can continue until it's largely accepted everywhere, like dollars are now.

Bitcoin also makes transactions less costly because they can be done without fees unlike credit card transactions. There a many reasons for merchants of all types to want to accept bitcoins.

EDIT: and the reason people want BTC (over dollars) is they can be used more effectively than dollars and may keep better, and even increasing, purchasing power; they can also be used more anonymously. With the recent price increases there are a lot of early Bitcoin adopters with valuable coins to spend. They want to spend them in stores and anywhere they can.

[Peter Todd]

Wouldn't it make more sense if some of that growing computer power where to go to make the network scale (make it faster and allowing it to handle more traffic)?

First of all you have to understand that the currently 25BTC reward that miners get per block, the inflation subsidy, is set in stone and just isn't going to change from the schedule Satoshi set. So miners are going to get, via inflation, a significant percentage of the total value of the Bitcoin market cap every year. Right now that percentage is 12%, and it drops to approximately 4% in 2017. That's a lot of money getting spent on mining.

Secondly hashing power is what ultimately keeps Bitcoin secure. A 51% attacker has to somehow get more hashing power than the rest of the Bitcoin network combined. Hopefully that is just too hard to do, but if we make it easy enough that someone can, Bitcoin has a very, very high chance of completely failing. So we can't take that risk lightly.

The thing is mining income doesn't only go to purchasing and operating mining equipment, it also goes to the overhead required to operate a full Bitcoin node that stores and processes transactions. Bitcoin has a limit of 1MB of data per block, or roughly 7 transactions per second. With that limit the cost to operate a Bitcoin node is pretty small, and the network connection required to do so doesn't need much bandwidth. Importantly, you can contribute to Bitcoin's security via mining behind anonymous internet connections easily, such as Tor.

If we put more transactions on the Bitcoin blockchain, well, miners are just going to have to spend more money on network connections, disk space, and servers, money that could have gone to the hashing power that keeps us secure. For Bitcoin to grow to the level of, say, Visa, you'd need to be buying a new harddrive every week or two just to store the transactions that haven't been spent yet, not to mention you'd need a very expensive data-center-level internet connection.

Finally, it's not like the computing power that goes to mining is general purpose. It's a very, very specific math problem that is most efficiently done by highly specialized processors with chips (ASICS) built to mine Bitcoins and nothing else. You just can't use that computing power for anything else, and if you could, Bitcoin wouldn't be secure because you need that math problem to have very specific properties, properties that mean it has to be useless for any other purpose.

[acoindr]

...
Finally, it's not like the computing power that goes to mining is general purpose. It's a very, very specific math problem that is most efficiently done by highly specialized processors with chips (ASICS) built to mine Bitcoins and nothing else. You just can't use that computing power for anything else, and if you could, Bitcoin wouldn't be secure because you need that math problem to have very specific properties, properties that mean it has to be useless for any other purpose.

Actually, you can do merged-mining for bitcoins and other alt-coins (not Litecoin which uses scrypt) and have Bitcoin be just as secure. Also, not everyone uses the same equipment, with the same dedication, for the same reasons. For example, it's still quite possible to mine bitcoins with a standard desktop computer. It's probably not very smart or efficient, but it's possible. Some people may mine for novelty or curiosity or fad or whatever else. Not everyone will use the most specialized and efficient methods, but every computation that goes into mining, at any level, adds that much more security.

[kwukduck]

If you were buying something for 1million dollars, 6 confirmations would probably be required.

For a small transaction however, how hard is it to fake a bitcoin transfer?

Would anyone attempt it to steal a coffee?

After 6 confirmations, the transaction is 100% safe.

What about after 5, or even after 1?

There is not 100%, but it becomes increasingly difficult to cheat as more blocks land on top of it.

Also do not mistake 6 ltc confirmations with 6 btc confirmations, 6 ltc confirmations is about as secure as 1-2 btc confirmations.. (yes yes, i know the ltc pumpers like to tell you differently.)

[acoindr]

...
Also do not mistake 6 ltc confirmations with 6 btc confirmations, 6 ltc confirmations is about as secure as 1-2 btc confirmations.. (yes yes, i know the ltc pumpers like to tell you differently.)

Make sure you tell the complete story.

The only thing which secures any cryptocurrency is the total hashing power of the network mining that coin. There was a time in Bitcoin's history when it had less security for 6 confirmations than Litecoin has now for only 2 (for example, when it was only Satoshi and Hal Finney mining BTC).

As the value of any cryptocurrency increases so does the interest in mining it, and therefore the security for it rises as more participants join in.