Topic: ASICs are good and more centralisation is inevitable (Read 1525 times)

unfortunately I have to agree:

network topology and brittleness aside, the optimum engineering solution, the solution that gives the highest hash per $ is one specialised datacenter, close to a power generation facility. of course it could not rest on its laurels for long as competition would evolve to undercut it.

i think in the future their will be a smaller number of competing large hashing operations. very large data centers.

but there is more to worry about than the trend toward centralisation... a fractional reserve like protocol will be created atop the base Bitcoin protocol where user debt becomes bank credit all over again.

still you will make good profit by holding. HODL!

ASICS costs won't break even even now. Unless of course you manufacture them and mine at cost using money made from previous ASICs that people pre-ordered. (Pre-order is required now if you want a competitive rig, and it's a very bad thing ). Check KNC out, sold a batch of good ASICS, then their customers found them starting to mess about with the newer 20nm rigs and move goalposts, after selling thousands on pre-order...and then build a huge farm of their own. So. If they make that 13k rig for 2k and it won't break even for buyers but profit nicely for the 2k they are spending, why would they sell ASICs anymore?

Most manufacturers would be better off mining than selling rigs, ones who already paid their costs and banked what they need.

It's not just a matter of hashrate with the next gen rigs either. Last I saw, the KNC Neptunes suck so much power that the average U.S. house can't run them without a rewire. They'd be fine here, but that isn't going to be a good thing for small miners in future if power consumption follows hashrates quickly. It's getting industrial

WOW! You've made at the forum of this post, please send a link to me?

I see your point but I think it's misguided. There will always be one man who has a lot of $ and is able to purchase more hashing power than everyone commenting on this thread combined. In order to compete with the rich guys we would have to centralize and pool up to have a chance at solving a block before him.

WTF retards? Are you forgetting which browser you're on?

Interesting analysis. I said many times in this forum that solar-powered mining rigs is the future. Of course ROI is terrible for now, but it will change soon. Another possibility is mini-nuclear powered mining rigs. I know this is utopia. But who knows?

There are some really new and innovative coins out there that are proof that ASIC's arn't what makes one coin safe. You can have 51% attack proof coin w/o ASIC, you can fight centralization of mining, you can have more privacy etc. It all can be done (and should be done) on software level.

at the rate difficulty is increasing and new stronger asic's being made. It might be safe to go with asic resistant coins.   

i think in the future their will be a smaller number of competing large hashing operations. very large data centers.

but there is more to worry about than the trend toward centralisation... a fractional reserve like protocol will be created atop the base Bitcoin protocol where user debt becomes bank credit all over again.

The miners have expenses¹:

   (aggregate mining costs per unit time) = (cost/hash)*(network hashrate),

and 51% of this number represent the price per unit time to run a 51% attack.

   (approximate cost per unit time to run 51% attack) = 0.51 * (cost/hash)*(network hashrate)

Miners earn their revenue from coinbase inflation and transaction fees:

  (aggregate miners' revenue per unit time) = (inflation rate + transaction fees)

The difference between revenue and expenses is the miner's profit:

   (aggregate miners' profit per unit time) = (inflation rate + transaction fees) - (cost/hash)*(network hashrate)

In an efficient and competitive market where mining equipment is commodified, aggregate profits will be small, such that:

   (cost/hash)*(network hashrate) ~= (inflation rate + transaction fees)

The approximate cost per unit time to run a 51% attack is therefore:

   (approximate cost per unit time to run 51% attack) = 0.51 * (inflation rate + transaction fees)

So, neither the network hashrate nor the cost per hash really matters in an efficient market.


¹The aggregate value electricity, labour, ammortized hardware costs, etc., are included in cost/hash

The problem with this is that there will be companies supplying the clubs, first several, until competition eliminates most and only a few will be left. Also clubs will become more sophisticated and more expensive so fewer countries can afford to buy them.
However you look at it; asics lead to centralisation and there is no way that can be good for the trust in cryptos. Sooner or later this will inevitably lead to an attack.

Bitcoin is in a relatively safe time period;

1 - cpus   At first relativly decentralised until botnets take over
2 - gpus   decentralised
3 - 1st stage asics   few available, centralised production, very centralised
>4 - 2nd stage asics  lots available, several producers, reasonably decentralised
5 - 3rd stage asics   few available (costs), few producers, centralised

Only current gen asics are represented in this list, prev generation not used for mining anymore (but maybe used for attacks?)
The remaining problem is pc's are getting out of fashion. Will there be pc's in 10 years? And low cost high perf graphic cards?
LTC asics will not do much, they only render litecoin obsolete but do no harm to trust in cryptos (i presume that if anything hurts crypto's trust it has to be done to bitcoin). Bitcoin is just years from being too centralised to be trusted, it is time to move on to the next generation of coins.
I am biased towards vertcoin (duh!), vertcoin will keep its stance towards asics and they will have to be prepared to switch its PoW to something else if asics should get developed for scrypt-n. What is missing is a framework to do just that with the least amount of hassle for its users.

ps Also don't discard the possibility of (perps in post above) starting a nice asic farm just to kill the network. It may seem far fetched but it can easily be done if you have the funds.

If you consider things carefully, you will quickly realize that ASICs do not make the network any more secure. Yes, it's quite true that they do allow the "good guys" to have a much higher hash rate by purchasing them. On the surface, this probably seems like it makes the network more secure. But there is a problem. The "bad guys" are able to purchase the same ASICs as the "good guys". They would be attacking with weapons of the same strength, security is not affected by whether we use ASICs, CPUs, or GPU's. Everybody has the ability to purchase the same devices, therefore the relative numbers don't really change, only the absolutes.

Let me give you a small analogy. You'll just formed a small country, and you have to defend it from intruders. No weapons have been invented yet, all you have to fight with is your fists, this is also the only weapon your enemy has. Okay, so now clubs get invented, you get all your people armed with clubs now so that your nation will be more secure. Here comes your enemy to invade, seems they've got some nice clubs too. Well, it seems like your security is not much better than it was before, doesn't it. Well, you just heard some encouraging news. Somebody is just invented a much more powerful weapon called a gun, get a few of these and those club wielding invaders won't have a chance. So you get yourself armed with these new guns, you feel very safe. Then the enemy invader show up again, they don't seem to have any clubs with them, what do you suppose they are carrying?

You understand the point of this little story? As long as the same technology is available to all sides, nobody gets an advantage. This is exactly how ASICs "improve" network security.

Maybe, but its correct. Bitcoin's net hashrate could easily be 0.0001% of what it is now if the sha256 algo wasn't so ridiculously simple to implement in silicon.
Read the rest of the line: the cost/hash is what matters.

Now that is just adorable.

Really?

I cannot believe what i am reading here! Why don't you all go back to fiat?

Centralisation = BAD! That was the idea behind bitcoin! (ok, last exclamation mark)

The network hashrate has absolutly nothing to do with the network security. The cost/hash has everything to do with network security. Asics evolve and soon only one or two companies will be able to produce them economically and there's your centralisation pre-programmed to happen.
Or, some (choose your enemy; chinese/russian/middle eastern) chipfab with energy to spare decides to pump out asic chips at 10000/day (at a cost of a few dollars each) and snafu the network? That would have happened already if bitcoin was as important as the dollar. With relatively little money you can make your own asics! What do you think used wafersteppers cost?

Once centralised your pretty coin is dead. Only a few points of attack simply means those points will get attacked, be it from governments or crooks.

There is only one enemy in the long term, and its centralisation. Any coin should do its utmost best to keep asics away, f&$kit, hard fork any coin under thread at the first sign of those %$#^^#%^^% showing up.

a nuclear power plant to run a mining operating would be intensely interesting

It will be more decentralized soon. Asics would become cheap, many people would be able to afford asics in the future. And I don't think that Moore's Law would be applicable forever there is a limit to it. Soon they cannot make more powerful with same size asics. But the price will reduce due to mass production.

ASICs may have actually saved Bitcoin, because without ASICs the overall Bitcoin network hashrate would've been vulnerable to botnets and rogue supercomputer usage. Because of ASICs, no general-purpose supercomputing installations have any chance of remotely mustering enough hashrate to threaten the network.

IMHO that attempt to flee from hardware that can flood the market with hashrate (and explode difficulty) is a moot point.
No matter if ASICs, GPUs or CPUs - those with the intention and the cash available will always find it easy to aquire or produce hardware to suit their need.

There is no escape from that, that possibility is an illusion.
As soon as there's sufficient incentive, you'll see dedicated hardware hitting the market earlier or later. There's nothing in the common technology market that cash can't buy or create.

Hardware of any type always costs money, hence the wealthy participants will always be able to throw significantly more hardware into the race - and so it begins again.

PS.
The Scrypt ASICs are already out there with the big guns already in develpment (this developent is just like SHA256).
And even if ther was a completely ASIC-proof coin - guess what, you'd run that coin on your 2 Desktop computers while others would employ entire server farms.
That type of "wealth inequality" is so deeply entrenched in our world, it will always show up as soon as there's a race to profits.

Interesting analysis. I said many times in this forum that solar-powered mining rigs is the future. Of course ROI is terrible for now, but it will change soon. Another possibility is mini-nuclear powered mining rigs. I know this is utopia. But who knows?