Vitalik Buterin of Ethereum has published a proof of concept for a sharding update of the Ethereum Blockchain .
Sharding is a way to solve issues with the scalability blockchain, meaning, how well the system can operate as its scale and workload increases. Ethereum uses a proof of work system, meaning that all transactions must be processed by all nodes in the system. This is the system used by the original cryptocurrency and is a tried and tested way of upholding the integrity of the blockchain, but it makes the network heavy and slow. Many solutions to this problem have been proposed and launched - larger blocks, child blockchains and the Lightning Network are just some examples.
Sharding, as the name implies, basically means that the blockchain will be fragmented, lightening the workload of validation nodes. At the beginning of 2018 the Ethereum Foundation launched two research and development programmes to develop sharding protocols, funded by subsidies of between $50,000 and $1 million.
The most recent proof of concept discusses a mechanism that can be "bolted on top of the current ethereum main chain".
Source: ethresear.ch
A proof of stake-powered beacon-chain will be tied to the main chain. The beacon will act as a heartbeat, issuing new shard-blocks every 2-8 seconds. These new blocks will randomly suggest a node to verify it, and that node will have the choice of creating a shard collation or not.
The verifying node is randomly chosen but the randomness is affected by that node's activity. This is the hallmark of the proof of stake consensus system. Apart from lightening the workload of the network, it has the added advantage of ensuring its stability because nodes with higher stakes, and thus more incentive to maintain the system, have more chance of being selected to verify transactions.
It should be noted that each new block, of any type, will be dependent on a block from the main chain. However the main blockchain does not personally verify every transaction - they are verified by proxy; that is, by randomly selected validators.
Buterin adds: "A not yet fully solved challenge is determining how to incentivize and when to allow cross-links."
