Scenario of centralization of mining
In my last article I talked about the fact that Bitcoin is decentralized, i.e. no single point of failure, or vulnerabilities. Many critics have pointed out that mining to a certain extent centralized, and therefore, my arguments lose their force. In this article I explore in detail the centralization of mining, and will consider some scenarios to understand what are the risks, how it can turn and what are the implications of that.
The definition of the centralization of mining
Centralization of mining can have two values:
- Production of equipment for mining is mainly concentrated in the hands of a single company.
- A large part of computing power is controlled by a single company.
These are different things, and when they say “the centralization of mining”, it is not always clear what is meant. Next, the article examines the risks and possible attacks in each of these cases. Note that we will study here only some of the scenarios, but they should give a pretty good idea about the possible risks of centralizing mining, and their prevention.
The centralization of production
Bitmain really makes a great piece of equipment for mining dual with SHA-256 (a hashing algorithm with proof of work behind BTC, BCH and some other), and most of the computing power in the Bitcoin network at the time of writing these lines comes from the production of Bitmain miners. We suggest here that although Bitmain produces the majority of equipment for mining, the company does not necessarily control this equipment.
What are the risks of production of most of the equipment, ensuring network security, the only company?
Scenario 1: the backdoor
In this scenario, we assume that Bitmain sells the bulk of its equipment.
The risk is that Bitmain can implement in mining hardware a backdoor hidden by hardware or software. Here’s what backdoor:
- To make mining hardware to connect to a pool controlled Bitmain, Bitmain minasama selected value.
- To take control over the block template to list the rewards to addresses controlled by Bitmain.
- To remove effective evidence of completion of the work, if the miner does not refer to a pool controlled by Bitmain.
- Disable the miner with some predetermined signal (emergency stop).
The first two possible backdoor will be really obvious to anyone who is even slightly mindful. In addition, their existence is easy to prove, at least with a small log. The consequences of the discovery of the backdoor of this magnitude will actually destroy the reputation of the company Bitmain or, at least, will make her a defendant in collective lawsuits. It would be tantamount to a kamikaze attack on Bitcoin, which may hurt Bitcoin in the short term, but will completely destroy as a manufacturer Bitmain ASIC for Bitcoin.
The third and fourth may be less visible, but are still subject to discovery. Deleting evidence of execution of work units delayed, and pools that are not controlled by Bitmain seem to be unlucky. When using emergency shutdown equipment fails. The direct effect of both options is really hurting Bitmain, as companies need to understand the clients requirements to get the money back and complaining that miners are not working. Any benefits, such as attracting more people to the pool will be secondary. It is also associated with high risks, because competitors can detect these backdoors and use them to the detriment of Bitmain. One risk of detection destroys any reputation of Bitmain, and the advantages rather vague.
Note that this may be done with their equipment any manufacturer miners. The hardware is very difficult to audit, and buying equipment from a particular manufacturer, you are, in a sense, trust that he will not disappoint you.
4-5 years of existence, Bitmain has not resorted to such tactics, there is no reason to assume that ever will come running. The organization of such backdoors requires a lot of planning, however, high detection probability and/or failure and a low probability of reward sufficiently discourages the manufacturer.
Scenario 2: manufacturing defect
This scenario assumes all the same as in scenario 1, but the equipment has a critical defect. For example, the fire equipment when exceeding a certain temperature. Or if it incorrectly calculates the timestamp.
The worst thing there can be is when the equipment generates invalid blocks, and the rest of the network it is easy to see. Again, this is just hurting the manufacturer, since he will face complaints from customers.
Scenario 3: inflation of prices/limitation of purchase/delivery delays
In this scenario, the hardware manufacturer uses its dominant position to increase costs for buyers. We can talk about raising rates, mandatory use of certain payment methods, the delays, perhaps even about the limitations of how equipment can be used.
This tactic becomes intolerable in the presence of competition, as the total cost of equipment may not exceed the prices of competitors without sacrificing sales, and therefore it should be used with caution, if used at all. Additional income from such actions offset by long-term damage to reputation.
The centralization of computing power
Abuse of equipment manufacturers can lead to some bad consequences, but much more dangerous scenario is the concentration of computing power. In particular, if one company would control more than half of the computing power of the network.
Similar scenarios can be divided into two different categories:
- One company controls the pools, in total supplying >50% of network computing power.
- One company controls the cars, in total supplying >50% of network computing power.
A possible attack is similar, but the methods to prevent the attacks differ slightly. If a single entity controls the majority of pools, the pool participants can just go to another pool, to prevent the attack. If a single entity controls most of the machines, such a solution is untenable. Keep this in mind when we consider possible ways to attack the network entity in control of most of the computing power
Scenario 4: the blockchain most
The most obvious that can make control so much of computing power is to reject the blocks from everyone else, effectively appropriating all of the remuneration for the units. It may also reject not popular with his transaction and may try to implement a double spending attack.
Such an attack to carry out is not as easy as it may seem, based on only mathematics, if the majority is not much higher than 50%.
To illustrate why, imagine that a hypothetical manufacturer Mitbain controls 60% of the computing power and decides to carry out an attack deflection unit. The probability of finding the block the rest of the network is 40%. It is obvious that as the minority still has some processing power Mitbain at some point will lag behind the rest of the network to 1 unit.
To take the lead, Mitbain will need to find 2 units more than the rest of the network. It’s not as easy as it seems. If there is sufficient time and much of the computing power of the capture inevitable, but not necessarily will happen very quickly.
Here is some math involved, but the number of expected units to capture the rest of the network Mitbain actually quite large. At 60% of network computing power, the expected number of blocks to capture network Mitbain is 6! Note that at 60% processing power, so for these 6 units will take from 60 to 100 minutes.
In addition, in the best case scenario for the attacker, the whole network will cancel the previous 5 units to 6 new units for the attacker. All transactions held in the previous 5 units will be cancelled as if they never existed, and transactions of the new 6 units will be treated as canonical. It’s called a reorganization of the blocks, and it allows for Bitcoin to implement a double spending attack. Of course, the attacker may be “good” and to include more or less the same transaction that was captured in the original units, but there are no guarantees.
In such a scenario no rational seller or exchange will not accept a transaction with less than 30 confirmations (at least, without some knowledge about what’s going on).
The table above shows how many blocks can be reorganized every time the rest of the network finds a block. As you can see, even at 70% of network computing power to carry out this attack requires a lot of time. If, after spending enough computing power to search 6 blocks Mitbain will still lag behind by 2-3 units, will Mitbain to continue?
In addition, a major restructuring of the blocks signaled by the network, there is something dishonest, and nodes are likely to be suspicious to these new blocks. It is possible that the operators of full nodes in the network simply cancel these blocks (this can be done using the command invalidateblock) and will be happy to consider as the canonical blockchain, made up the rest of the network. Then Mitbain spend a huge amount of computing power, claiming their bad intentions and create a fork that is not recognized by many sites. It will be a hard fork without any protection from re-play, and the community will decide which blockchain more valuable.
In addition, during the attack, a major restructuring will occur every time the miner is not controlled Mitbain, finds a block, from-for what the payments would be extremely risky. In fact, if there is no control over approximately 80% of the network, during this attack the Bitcoin becomes practically unusable.
If Mitbain just mine will be fine on the network, the reward for mining will be virtually the same without any suspicion or damage to reputation. The cost of double-spending, theft of the commissions and the deviation of the transaction must outweigh the risk of failure, including loss of remuneration for mining, loss of reputation, and harm to Bitcoin.
Simply put, this attack is virtually meaningless from an economic point of view, because the benefit for the attacker is simply insufficient. Moreover, even in case of its success, Bitcoin will still survive! There is no guarantee that the temporary degradation of the network enough to all holders sold their bitcoins.
Scenario 5: disable computing power
However, to influence the greater part of the computing power is not necessarily to attack the network. The majority can just drop mine and make it proportional to network security. For example, most, say 80%, can give mine to exert political pressure for the introduction of a specific option. This will remind strike.
Such a scenario would cause some difficulties in the network. 10 minute blocks will turn into 50 minutes. Mempool probably very soon filled, and the transaction will take place very slowly. This, in turn, can lead to increased commissions.
However, such a scenario will cost the attackers much more expensive. They refuse 1437 BTC and at the current rate of approximately $10 million in revenue a day. Even if the equipment will be used in another network, it will greatly affect the profitability and costs of lost profits.
Moreover, the time of confirmation in the Bitcoin network in a matter of weeks to recover, while the lost income of the strikers will not return.
That can make a large amount of processing power
Perhaps this will be useful to briefly talk about the fact that a large portion of processing power to do not can.
First, the majority cannot take away from you the coins that you already own. All your coins are yours, and the worst that can an attacker do is to perform a double spending of inbound transactions or some time to obstruct the passage of your outgoing transactions. This is normal and expected, as we have seen in the case of a transaction with enough big Commission in December.
Second, the majority cannot change the rules of Bitcoin. In a sense, it can create a new rule of consensus, but it will be a hard fork which requires all updated. It can try to convince the rest of the network, its rules better, but as sovereign individuals, Bitcoin users are not required to adhere to such rules. The owner of the site is completely free to decide what rules to follow.
Thirdly, most can’t hurt you without hurting, to some extent, myself. Such an attacker could harm the network, but not without opportunity costs for themselves. He may try to perform a double spending, but not without significant risk to be blacklisted by many sites.
At least from the point of view of production risk is constantly decreasing. How would all either hated Bitmain, let’s not forget, what were the producers that existed before her. Butterfly Labs, CoinTerra and KnC Mining – just some of the names in this space, and they had serious problems even with doing pre-orders.
Bitmain has brought the mining industry professionalism, which before there was not simple. The company sold fully assembled ready miners at conferences in 2014, when the other delayed delivery on pre-orders made months and years ago. Competence, introduced by Bitmain in the mining industry – the reason for the bankruptcy of these other players.
However, there is no reason to believe that market domination Bitmain will be eternal.
First, there is a great competition. I know at least 4 start-UPS that enter into the space mining in an attempt to overthrow Bitmain. There are also large companies such as Samsung, Intel and NVIDIA are thinking about how to enter this very lucrative industry.
Second, unlike the Bitcoin, in the production of equipment for mining there is no strong network effects. Perhaps people wanting mine, a little worried about the manufacturer of cars but most are more concerned with how much money you can earn. In other words, if they are used to buy Bitmain miners, it does not mean that they will continue to buy the products of this company. On the contrary, many people who want to mine, are willing to pay more to get products not manufactured by Bitmain.
Thirdly, now Bitmain is a very big company. She is working on ASIC for machine learning, ASIC for alithinou, repurchases of the company and finances many different projects. Large companies are often less flexible than their smaller competitors, and eventually such large companies it is possible to detect flaws.
This does not mean that Bitmain will just abandon their large share in this very profitable industry, but, definitely, there is a large space for competition. If you, like me, believe in the free market, it is easy to see that in the long term, any imbalances will be smoothed out. Now there is an imbalance of production. The centralization of production equipment for mining is a short – term problem.
Centralization of mining has long been a bogeyman for members of the Bitcoin community. Worse yet, many still believe that the majority can “control network”. Independent properties of decentralization a lot of help, and Bitcoin is much better protected from such centralized control than many people think.
In addition, centralization of mining is not structured to take too long. Mining is a commodity game, where prices typically decline. Obviously, Bitmain will do everything possible to protect existing market share, but such attempts without producing a better product are usually expensive and short-lived.
And if mining is no single point of failure, Bitcoin will survive.