Module jormungandr::leadership

new module to prepare for the new leadership scheduling of blocks

here we need to take into consideration that won’t have access to the cryptographic objects of the leader: they will be executed in a secure enclave.

data structures

We need to separate our components as following:

1. the enclave:
   • upon receiving the necessary parameters, it will return a schedule when it should be elected to create a block;
   • upon receiving the necessary parameters, it will create the proof to finalize the creation of a block;
2. the schedule:
   • it holds the schedules for a given epoch
   • we can query it to get a list of schedule for the REST API (useful to have information when the node is expected to create blocks);
   • optional but useful: have a way to update if a schedule has been executed (and what time);
   • optional: have a way for the blockchain task to update the schedule to know if the scheduled block as been accepted in the branch;

The enclave is not yet implemented, but we will need to separate the crypto from the representation here.

workflow

The flow process will work as follow:

1. the leadership module will receive a new event to create prepare a leadership schedule; It will only includes the Leadership object from chain_lib and the TimeFrame active for the future blocks to come;
2. upon receiving these data, it will query the enclave to know the list of expected scheduled leader elections; (this part may require heavy cryptographic computation, we may want to split this part into incrementally long queries);
3. once the schedule is retrieved (even partially) we can start waiting for the appropriate time to create a new block (to run block fragment selection) and ask the enclave to sign the block;
4. once a block is ready we need to send it to the blockchain task to process it and update the blockchain.

how and when to trigger a new leadership event

The blockchain module has the material to create the new leadership parameters for a given epoch (the Leadership object and the time frame). It needs to send the appropriate data when necessary.

2 ways to trigger a new leadership schedule from the blockchain module:

1. the blockchain detects an epoch transition,
2. the leadership sent an end of epoch signal to the blockchain;

Now doing so we may trigger the same leader schedule twice. We will need to make sure we don’t duplicate the work everywhere.

Re-exports

• pub use self::enclave::LeaderEvent;

Modules

• enclave 🔒
• logs 🔒
• process 🔒

Structs

• Enclave
  represent the client side of an enclave. From there we will query the actual enclave about schedules and signing blocks
• LeadershipLogHandle
  leadership log handle. will allow to update the status of the log without having to hold the Logs
• Logs
  all leadership logs, allow for following up on the different entity of the blockchain
• Module
• ModuleConfig

Enums

• EnclaveError