3 ms·
About trust: You could distribute the load randomly, and do double work, if two data sets doesn't give the same output you do the calculation again and blacklis
by z3t4 9y ago
About trust: You could distribute the load randomly, and do double work, if two data sets doesn't give the same output you do the calculation again and blacklist the node that was wrong.
- rbanffy 9y agoYes, but you won't want any sensitive data to be processed on the platform unless an irreversibly unrecognizable transformed version of it is what is processed.