Blockchain is something that requires long-term exposure to learning, so this may also become a series. Learning has just started, and I will share some experiences regularly, as long as I can make progress.
A brief look at the underlying technology of the blockchain can be seen as a distributed system, so all distributed systems cannot avoid the CAP principle. The design of the blockchain abandons the consistency in CAP and adopts eventual consistency.
This consistency is similar to many distributed coordination middleware algorithms, which can be basically understood as a variant of another Paxos algorithm, which is not unusual.
I said before that the core of the blockchain is: distributed, decentralized, irreversible, and tamper-proof.
In fact, decentralization should be called weak centralization. Decentralization is all individual nodes at the bottom of the technology. There is no concept of master. Cassandra and redis cluster have the concept of decentralization. To synchronize data through rumors, P2P communication should be An RPC call, these two pieces are not too unusual.
But the reason why it is not a complete decentralization is mainly considering that the future application scenarios of the blockchain are not only the underlying technology, but also need to be supported by certain real life scenarios, so in real scenarios, if there is no powerful It is difficult for a centralized organization to do endorsement, and it is difficult to gain everyone's trust, so this centralized organization plays a certain endorsement role.
In fact, I am more interested in how the blockchain can achieve irreversible and tamper-proof.
Basically, it seems to be a combination of mathematical knowledge plus computer cryptography knowledge, combined with Hash method, plus random numbers to achieve a certain random purpose, plus timestamps can be tamper-proof.
For example, every 10 minutes, the entire network will conduct a hash computing competition with a certain degree of difficulty. Only a node that submits its own hash result and passes the network verification will join the new chain. This process is called proof of work.
The verification process is through Merkle tree verification.
As far as the motivation of game theory is concerned, it is mainly based on the strategy that the cost of tampering is far greater than the benefit. The cost is to rebuild a longer blockchain and require more computing power.