3.3 KiB
This system is meant to represent a body of experts receiving fees to perform work. Experts register their availability to receive work via the availability contract. Request and associated fees are sumbitted via the business contract. Evidence of the work performed is submitted as a post to the forum. A successful validation pool ratifies the post. Reputation is minted and distributed. Fees are distributed.
What if we want the work to be block production for a blockchain? Then to perform this work, an expert will need to participate in a communications network such that they can confidently arrive at a majority view of each block. Or else must at least be able to attest that a proposed block is valid, meaning that it
- does not conflict with what the node believes to be the majority view
- includes what the node believes must be included according to the majority view note that with this scheme there be muliple possible valid proposed blocks.
In any case, block production will require some form of consensus protocol. (BFT).
We have to define an algorithm from the perspective of a single node.
That node will need to make certain assumptions. Let us identify those assumptions.
- continuity guarantees?
- sender identifiability?
- it needs to bo possible to verify an asymmetric signature
This leads us to the storage requirements for a node.
For a node to exist, it must be capable of at least some temporal continuity. That's what distinguishes a node from a client.
We want our protocol to involve performing certain kinds of work.
- Block production
- Messaging protocol
- In-memory storage
- Queryable history
- Fast record storage
- Archival record storage
What is common among these?
- Availability stake represents commitment to perform the specified work
- Peers must validate the work product via validation pool
How can we adapt the following concepts?
- Business contract interfacing with availability contract
- Reputation is minted via validation pools
- Reputation is rewarded to validation pool winners
- Reputation is awarded to a post in the forum
- Reputation is propagated via citations
In a messaging system, the work is
- Listening for messages
- Receiving messages
- Processing messages
- Sending messages
- Maintaining context related to incoming messages for the duration of some operation
- Performing computations
The work of verifying peers in a messaging system is
- Detecting invalid messages
- Successfully defending against DoS attacks
- Initiating validation pools?
- Voting in validation pools?
The work of providing a storage service extends that of participating in a messaging system.
- Storing data
- Retrieving data
The work of verifying peers work products in a storage network is
- Periodically querying peers and verifying their responses
- Participating in validation pools to police peers
- Initiating validation pools
- Voting in validation pools
Governance of a storage network includes tuning post and validation pool timing and other parameters. This can be served via the forum and validation pool, by having the clients agree on an interpretation of the forum, such that clients can derive from forum posts, at least some operating parameters. It may even be possible to use the forum to provide the client code itself, or tools for generating such code.