With cryptocurrencies, there is always a question of how the blockchain will scale as technology changes and the currency grows in demand.
Blockchains are meant to be immutable, meaning that once a change has been made to it, such as the data for a particular transaction, the record of the transaction cannot be changed or forgotten. This means that, for one thing, the distributed ledger that holds the record of all the transactions will inevitably get larger and larger, and any computer that wishes to be a node may have to download a potential cumbersome file with all that data.
This is one of the things Ethereum has made efforts to sidestep. Computers in the network can take on the role of nodes with different tiers. Only “full nodes” have to download the entire blockchain history. The tiers in-between might download only the most recent 1500 transactions, for instance, since having a history to fact-check is crucial to a blockchain, both in terms of verifying that a node is legitimate and present on the blockchain at a specific point in the history of the blockchain and in terms of verifying the transaction data to follow. Transactions and computations on the Ethereum network might have a timestamp, but they must be somehow organized among a sea of other data seeking validation in the context of the current sea data where computers cannot all synchronize their clocks in a reliable manner, be sorted into blocks within it and verified using proof-of-work or proof-of-stake that prevents double-spending, centralization and collusion among the nodes, and so forth.
One of the unique ways that Ethereum characterizes blocks is as “uncles,” which would only be orphaned, too-little too-late blocks on the Bitcoin blockchain, and the uncle blocks serve to reward miners who were not able to be the most competitive at the time, which prevents centralization and collusion between the nodes, and it provides a unique color-characterization to the history of the blockchain that allows the lighter nodes, which do have the full, final blockchain history, to participate using part of the uncle-history as well.
While the finer details of how this works easily escapes grasps, it is part of how Ethereum is able to reliably validate blocks in under 10 seconds instead of 10 minutes, which is the time Bitcoin’s proof-of-work algorithm takes. This does help Ethereum maintain the speed to scale with increased network traffic, which has caused bottle-necking for Bitcoin, but Ethereum is planning to do away with proof-of-work in the near future. When it does, there will no longer be miners, only validators who will use proof-of-stake instead.
Proof-of-stake does not require the time or electricity that proof-of-work mining does. The trustless validation is designed to be maintained by requiring that validators have more to lose than they could gain by feeding back information to the network. There would be a buy-in where a certain amount of their Ether would be held in an account that would be forfeited if the node was caught trying to cheat the system. Some altcoins already use proof-of-stake, so Ethereum would not be going into it blind and the transition has been planned out since the founding of the Ethereum Project. As with other open-source cryptocurrencies, the community and the stakeholders of Ethereum can choose down the road if changes are needed to allow the network to scale, be it by forking or just updates, but it would appear that the platform has been designed well enough to scale seamlessly into the future.