Solving the Nothing at Stake Problem
The Proof-of-Stake (PoS) model has a significant flaw. Validators can start supporting multiple blocks at the same time to increase earnings. In this case, they risk nothing and receive a reward in any case – this problem is called “Nothing at Stake”. In PoW, this issue is solved due to the high cost of the equipment and electricity used: it is unprofitable for miners to invest computing power in short branches. In a virtual PoS environment that does not use physical resources, the validator does not care: it does not incur any tangible costs by placing bets on competing blocks. The Casper protocol is designed to solve this problem. It introduces a penalty for maintaining a short branch. A validator who has placed bets on two blocks loses his money.
The Byzantine mistake solution
The second serious problem, the solution of which is laid in Casper, is the Byzantine mistake – a situation where some of the participants take actions that impede the achievement of consensus and the operation of the network. In PoS, at least ⅔ of all registered validators must vote for a block entry. When a certain number of inspectors go offline, this ratio cannot be achieved, and the network is blocked. The algorithm used in Casper allows such validators to be disconnected from the network, and their rates are lost.
Improving network scalability is achieved through the use of a mechanism for splitting the complete blockchain ledger into shards. Casper eliminates the possibility of economic scaling. It means that businesses with large operating activities gain a competitive advantage over others by lowering cost per unit of output.
The move to virtual resources will ultimately eliminate miners and stop the gigantic power consumption. The need for expensive highly specialized and powerful computing equipment (ASIC miners) will disappear. This will increase the income of participants involved in checking and writing blocks.
Avoiding expensive equipment
In order to become a validator and make money on block writes, you do not need to buy equipment subject to wear and tear, breakdown and technological obsolescence.
Casper protects the network from dangerous threats in the form of 51% attack, centralized control, and malicious actions of validators.
The model is based on performing complex mathematical calculations that require large computing power. Three main problems of PoW arise from this:
It was to eliminate these risks that the Ethereum developers decided to switch to another method of reaching consensus – Proof-of-Stake (PoS).
The scenario of Casper Friendly Finality Gadget implementation is accepted as a working one. It implies a hybrid coexistence of PoW and PoS protocols at the first stage of the transition.
The essence of Casper FFG is that the overwhelming majority of blocks are achieved through proof of work. But every 50 blocks, a checkpoint is created at which the block is recorded using the PoS algorithm. This block is called a key, or checkpoint.
At this point, the concept of “finality” (“completeness”) is introduced, meaning that the validators fix the given state of the blockchain, it cannot be canceled or corrected in any way, even if 99% of miners begin to maintain a chain that did not contain this block.
There is only one option for returning a completed block. It is called slashing, or "cutting conditions". This mechanism says that no matter how such a situation arose, there must be such a set of validators, with a total of at least ⅓ of the bids from all voting participants for the completion of this block, who send messages to revoke the completion of the block.
In this case, the entire validator's deposit in the Casper's contract is destroyed. Except 4%, which goes to the validator who sent the “proof” of the need to enable the cutting condition (“seeker fee”).
Returning a completed block is very expensive for validators, and this serves as a defense against returning completed blocks.
The introduction of the hybrid model is due to the simplification of the transition to the new mechanism for both participants and the system.
To speed up the work with the registry, Casper developers propose to use a technology from the processing of large databases – sharding (shards). According to this method, large tables can be divided horizontally into several small ones without losing data and structure, which significantly speeds up the speed of working with information.
Casper eliminates the possibility of economic scaling. It means that businesses with large operating activities gain a competitive advantage over others by lowering cost per unit of output.
Proof of ownership is often criticized for making the rich even richer: the rewards are proportional to the rates, which means that the higher the rates, the greater the income.
The problem of “Nothing at Stack” in Casper is solved by introducing penalties in the form of burning out the validator's bet. But a situation may arise that at first he chose one block, and then switched to another, realizing that this block is correct. There was no malicious intent in the validator's actions, but he will also be punished.
One of the solutions to this problem is the introduction of several rounds of voting. The consequence of such approaches is another task that developers need to solve – a constant increase in logical and protocol complexity.