Watching a well produced blockchain video I still say bullsh*t. This is all about providence but … My issues include:
Having been in database application development for many years, including fentech, I’ve had to use
all manner of database tools… Oracle, SQL Server, BDB, SQLite, mysql, postgres, redis and so on.
Replication, high availability and replacation are still a serious challenge. In recent years
RAFT and a few other protocols have made replication easier to reason and while I’ve had to deal
with database state and consistency record
UPDATES are the most difficult to replicate and
updates that effect indexes can have very nasty side effects.
The most important failure in blockchain is what to do when it fails or someone does make a change to the data or underlying structure either intentionally or unintentionally? It’s important to detect failure conditions but then how do you repair it? Repairing a damaged blockchain would seem to be an antipattern.
One problem with ledger design is the amount of inactive data it produces when doing pseudo updates.
While the blockchain is still BS it does have some silver lining. The ledger approach to schema design makes sense although unexplored. The blockchain does not implement indexes so there is some added work but there is no pruning only re-balancing. Appending records to the DB is faster than updating which also reduces the “work”.