What is sharding? Explain the solution the scalability of Ethereum

Imagine that you are sitting in the great hall and write the exam, along with another 500 people. For example, 50 teachers observe and evaluate the exams. Imagine that each of these 500 examinees will need to pass all 50 teachers and your grade will be the average of all exposed. Can you imagine how much time it will take to write all the examination papers? Obviously, this is never a practical method, which only complicates the exams and becomes more difficult with increase in the number of students taking the exams. Exactly the same challenges that faced the Ethereum blockchain, gaining popularity and increasing the volume of transactions.

Now imagine that instead of 50 teachers who need to make 500 exams, the responsibility will be divided into segments. Each inspector of the teachers would be responsible for 10 exams and assessment, which exposed every inspector would be final for each of these 10 students. If the inspector is not sure how to rate a particular student, he can consult with other inspectors and to obtain more information.

Actually, this is a simple example of how sharding can solve the problem of blockchain Ethereum, increasing transaction speed and reducing network load.

scalability Problems of blockchain Ethereum

Currently, the blockchain Ethereum requires that all the nodes in the network stored and processed through the transaction. In other words, the problem of blockchain Ethereum is that he needs full nodes (nodes). Although this process ensures the network a high degree of protection because each transaction must confirm each node is very complex and poorly scalable method, which the developer community Ethereum plans to stop.

The greater the number of nodes, the slower and less scalable it becomes the blockchain.

In its current form the blockchain Ethereum needs to make a choice to meet two of the three following attributes:

the
1. Security
the
2. Scalability
the
3. Decentralization

The Current state of blockchain Ethereum fully meets 1 and 3 criteria, but is very lose when it comes to scalability. But times change, and scalability stone lies on the way to a massive expansion of the blockchain, so the developer community Ethereum decides to partially sacrifice safety. So they hope to increase the speed of transactions.

Solution: sharding

Imagine that Ethereum is divided into thousands of Islands. Each island lives its own life. Each island has its own unique features and all that this island belongs, that is, the accounts can interact with each other and free to use all the features of the island. If they want to reach other Islands, will have to use a specific Protocol.

In the usual sense of sharding is a kind of partition which divides a large database into smaller groupings and faster, the shards (shards). By definition, the shard, aka the shard, it's part of something whole.

But in relation to the blockchain Ethereum is a very complicated and complex concept. In this article you will find a simplified explanation, but for a detailed study, you have to immerse yourself in the many technical arguments and mathematical calculations, of course. In short: the network is divided into many smaller sections shards and each shard contains an independent part of the transaction data. The Ethereum community is trying to get away from the need for full nodes due to the introduction of sharding. Sharding will be placed solely on the Protocol level.

Implementing sharding will allow each node to store a subset of the data and be responsible for the verification of a specific set of transactions, not all transactions that are consistently on the blockchain due to non-parallel nature of the Ethereum Virtual Machine (EVM).

When a particular node needs information that is not stored in its own block, it will find another host that has the necessary information. This precondition for cross-sadovyh messages.

Currently, the blockchain Ethereum allows for 8 transactions per second (TPS), and when sharding will be available to thousands of TPS without the need for full nodes, and the total size of the node also reduced. However, this process cannot be considered completely safe, because nodes are more likely to be dependent than independent. The sacrifice is ready to go the developers to increase the network bandwidth.

Why Proof-of-Stake is required for sharding

All of the above, this risk will only take place if Ethereum will continue to work on the mechanism of consensus Proof-of-Work. Since PoW relies on the hashing power of the network, there must be a way to prevent the attacker attempting to manipulate a hash the network, which will be much lower due to partitioning of the network to shards.

Attack of 51% of the blockchain for PoW database without sharding practically impossible because of the power required by the hash, and the total cost of this attack is that in the case of Ethereum will result in billions of dollars. The attacker would need 51% of the total power hash of the network to attack the network and to control a majority of votes.

A blockchain-based PoW with sharding the probability of attack is 51% higher and the overall threat to the security of the network above. An attacker or group of them can concentrate the power of the hash to a single shard and get control of this shard. The attacker can then use the cross-Sadovoy communication to attack other shards in the network.

It is that EVM was safedecentralized and scalable. PoS would Ethereum to deprive the attacker the ability to concentrate hash rate on the same shard. The PoS system would also eliminate vulnerability to attack 51%, which would be more relevant if sharding was implemented into the system on the basis of PoW.

PoS blockchain with sharding will deprive attackers of the ability to choose the shard that they want to process, and will not allow them to know in advance which shard will be assigned as decisive (authoritative) in the network.