Future Protocol Enhancements

TQuorum was a great place to meet all the people involved in the development of Tezos. One of the most interesting parts to me was the discussions I had with Adrian Brink. A few points we discussed

  1. Implementing Ouroboros algorithm. This would, if I understood correctly, not reveal which baker would be due to bake/endorse until the very moment the operation happens. This would be a real improvement in protecting bakers from premeditated attacks.

  2. The use of something similar to Polkadot’s Grandpa & Babe consensus mechanism. I had previously mentioned this on reddit and stackexchange but it was not discussed much. Reddit Post

  3. An algorithm, that allows wannabe bakers to offer themselves as bakers to potential bond pool partners which would automatically assign bond pool members to the baker until they reach at least 1 roll.

  4. Bond pool partners who keep their bond pool and don’t have to send it to the baker physically. Their bond would then be frozen proportionately to their share of the entier bond pool and their rewards distributed equally so. This would solve a lot of the legal issues for bakers who use bond pools and take physical delivery (e.g. TzDutch )


Regarding the first point, the specific part in Ouroboros which allows for dynamic leader election is where a participant uses a VRF (in Praos) and a randomness beacon (derived from prior blocks / time) to prove that they are the block producer for that slot. Another example of where VRFs are used for leader election is in Algorand. VRFs can be a good solution to picking a leader and it is something that is under discussion in the consensus team at NL as a possible future improvement.


Thanks for information