Blocks in a blockchain protocol have several functions. Major function is clearly to have an order of the transactions, which order is the same for all of the distributed - decentralized nodes in order to avoid double spending. Like if there a two transactions that are double spends, every node can consider the first one as valid and the second one as invalid. Just for avoiding double spending however, we do not need an order. A partial order of transactions is enough in a sense if two transactions want to consume the same account or transaction there must be a common order among these two transactions. So, instead of blocks and full orders, it clearly makes sense to investigate more efficient storage structures for transaction processing.
