Why a public blockchain ruin your startup?
My post about the economy of the blockchain caused some discussion, but I saw that most of its participants did not see the central thesis, looking at the problem with a purely engineering eye. The problem is not engineering, but only an economic nature and I look at it as an entrepreneur who faces a choice:
What is expedient — to use someone else’s blockchain for your project or to make your own? In what situations is it advantageous to use someone else’s blockchain, and under what do you do?
I’m discussing a bit and invite everyone to participate.
A blockchain is an ordinary database like regular electronic tablet or a diary where you write your business, but there are two differences:
- You copy all the records on several computers, i.e., you immediately have participants, each of which has a copy, and where more than one participant appears the network.
- You agree in advance with all participants about how you overcome the errors that can occur when copying records.
This so-called consensus problem, which arises whenever a group of people has to come to an unambiguous understanding of the situation. If everyone agreed — a consensus has been reached, but if they did not agree — the protocol continues until a consensus is reached.
The protocol is just the negotiation process to order to consensus where every participant follow the common rules.
A good protocol leads to a QUICK consensus, i.e., it has SMALL COSTS for its maintenance, in the limit close to 0, but always above 0.
A bad protocol may not lead to consensus at all, i.e., it can have theoretically INFINITE costs for their services since negotiators must constantly eat, sleep and shit, burn energy in the negotiation process.
Therefore, the protocol as a kind of process always has costs, because it is simply a consequence of the law of conservation of energy. Thus, ALWAYS there is a protocol cost that will service any request and reach any consensus. And these costs depend on what protocol architecture, how many participants are involved in it.
The topology of the protocol architecture will always lie between the two extreme states:
- 100% centralized star — classic client-server architecture.
- 100% decentralized p2p topology, which as a result of the difference in the processing resources, which in reality will always be a kind of scale-free network.
Any architecture will have ONLY ONE KEY METRIC — the rate of decline in the transaction price as their volume increases. That’s exactly what the engineers ignore.
There are no performance problems; there is a problem with the cost of processing transactions because the blockchain protocols are designed to reduce the level of errors in the system since fraud is just a kind of error in the system.
So far, what I see in any existing blockchain protocols is the idea of reward for processing, which is logical. In this sense, any service takes money for processing requests.
In classical client-server systems, the cost of transactions FALLS with the increase in the number of transactions, but in the blockchain, the cost of transactions increases with the number of transactions.
The reason is that the more transactions, the more the node earns. Therefore, it is advantageous to open new nodes, so people are rushing into this business. Accordingly, the margin falls, mining becomes less profitable, any new node is a competitor. It means that the ratio of the number of nodes and transactions is easily calculated when any mining becomes unprofitable.
To prevent the network from getting up, you have to raise the price of transaction processing. And this process is endless. This is a systemic dead end. While average transaction checks are high, this is not sensitive, but large amounts of transactions are typical for trading.
However, 99% of transactions in the world are micro data exchange transactions, in which actual financial transactions can occupy hundredths of a percent. Even if we assume that no more than 1% of data-transactions are expedient to process through the blockchain, it still does not make economic sense.
Many people have written me, that way to process all transactions if we can to put in blockchain only every 10th or 100th instead?
I laughed for a long time. Ignorance is booming))) This is an old topic about the sampling frequency and the Kotelnikov (Nyquist-Shannon) theorem only in a broader sense.
Practically it was applied in Analog Digital Conversion where, for example, you can use 4–8 or 12-bit sampling frequency. Obviously, 12 —bit conversion has high accuracy, but if you want to convert quickly, the chip starts consuming a crazy amount of energy and releases heat, that means — you need special heat exchange mode, coolers, price. Therefore, we have to make compromises.
So, the amount of data grows exponentially.
Even if you put the number of only representative transactions in the blockchain which meets the all your requirements and Kotelnikov (Nyquist-Shannon) theorem requirements, it’s early or late the number of even only key transactions will lead to an increase in value.
All modern storage and data transfer systems are faced with the problem of their growth, and all solutions are directed not so much to increase productivity, but to lower the price of storage/transmission / processing.
In this approach, there is an economic meaning. But so far I do not see that in a public blockchain. While the cryptocurrency is speculatively increasing, everyone ignores this factor, but the period of exponential growth will move to the plateau. And what happens is that with 90% of retailers and real estate in 2008. While the market was growing, nobody cares about the cost of servicing loans.
As soon as the turn occurred, the cost of maintenance of transaction costs (including crediting) killed a lot of inefficient companies. As long as the bubble grows, everything is fine, as soon as the speed slows down or correction, the end comes.
Anyone who makes a bet on a profit model that is associated exclusively with growth will fail. Only those who create real services will survive, even with zero growth. But while the HYIP — it is necessary to earn it, so ICO !!!!
Some people may think that these reflections are something against the blockchain. I just saw a problem that needs to be solved. Obviously, if the solution leads to an increase in direct costs with an increase in the number of transactions — this is a vicious decision in principle.
Look at the WP of majority of ICO projects.If financial model doesn’t take into account the price of processing and the growing the price with the increase in the number of transactions,you see the high risk.
Any project will always have a choice to create own blockchain or to use the public. In public, there is more liquidity and people, but the price of these leads may be outside the economy. When we create own private blockchain with a limited number of nodes, the costs of attracting leads remain, but the cost of processing becomes acceptable.
I looked at the cost of bitcoin transactions to check my reasoning:
If your transaction size is less than 11–12%, then its price is about 10%. When a volume of transaction increases, the commission falls. And I decided to send a micro transaction, only 1 dollar. It took me $ 2.23)))
