Any computation, considered as a sequence of steps, can be described by a SHA-256 hash.
Therefore, we can modify our consensus algorithms by allowing anybody to join a subnet as an additional node that will check the SHA-256, provided that he has access to WASM codes and messages.
We can then somehow reward these additional nodes for upholding security better. This way, we probably can somehow get rid of distinct fiduciary subnets, etc.
Good idea? Any novel thoughts in it? Viable to implement?
I want to ensure that my social network state is not damaged by 7 malicious of 13 nodes.
So, I create an “additional” (having less rights than 13 nodes) node on AWS that keeps all the state. If 7 malicious nodes choose to damage the state, I can restore it.
Two comments. First, this would change the current ‘private by default’ design. When everyone is allowed to verify the block contents then you (accidentally?) remove any privacy the canisters currently have. This is desirable for some applications, but with only public data a lot of apps become either useless or require major redesigns
Second, this will only allow detecting rogue subnets, restoring is not fully possible. If a full subnet goes rogue it’s impossible to restore e.g. the subnet’s public key