I complete understand the database problem on blockchains... it is something that has to evolve otherwise they are not able to become industrialized quickly.

For me, and being part of high performance computing market (where most of things have to consider levels of second/millisecond), it makes me beleive that the next step is tiering on databases... so that, for example neworks become a collection of nodes that are not ALL with the same database. This already "somehow" exists but in a different concept and for other purpouses, for example, on EOS for the historical indexed information that records all the blockchain activity.

Thanks for staying around.