Sign the hash of the credential, not the raw credential

[noandrea]

In the code for signing credentials here:

elesto/x/credential/client/cli/tx.go

Line 355 in c2e33f6

func sign(

and in the code to verify credential signatures here:

elesto/x/credential/credential.go

Line 202 in c2e33f6

if !pk.VerifySignature(wcCopy.GetBytes(), sig) {

we do sign and verify the entire credential byte slice. The signing is done on the client side so it's relatively not critical, but the signature verification is done on the node side.

The signing computes the signature on the whole byte slice of the minified and encoded JSON. That seems to imply that the cost of signing/verifying a credential is linear with the credential size. The task would be to collect empirical data on the cost of signing/verifying a message vs computing the hash of the message and then signing/verifying the hash.

As for hashing algorithm consider blake2b or sha-3

[github-actions]

This issue was marked as stale because there was no activity for 15 days