There's decades of academic work in the form of "impossibility" results showing that handling byzantine fault tolerance in distributed consensus is alot more co

There's decades of academic work in the form of "impossibility" results showing that handling byzantine fault tolerance in distributed consensus is alot more computationally expensive than omission/crash fault tolerance. The following (FLP impossibility result) being one of the most famous ones: https://groups.csail.mit.edu/tds/papers/Lynch/jacm85.pdf https://groups.csail.mit.edu/tds/papers/Lynch/jacm85.pdf