# Overall Architecture

## Monetization Layer

- Player-data-as-NFT
- Player resources bound to the server they're produced on as "localized" unless "exported", fees incurred
- F2P means only ever being on one server.
- but if global subscription, server owners get % of the fee as rewards based on stake
- "export certificate" on goods means paying into the overall network
- server hosts gain % of export fee back as reward, based on stake
- servers can only export resources per epoch based on overall stake
- Players can sell "export certificates" for IGC (ingame currency)
- Players purchase "export certificates" with RWC (real-world currency; fiat/crypto).

## Servers

- Finite State Machine-based game with a "RNG Oracle" zero-knowledge service providing randomness.
- Oracle would need to have reproducible sessions (i.e. allow for post-date verification of gamestate & random numbers used)
- Could be completely isolated or publish secret nonce information after an epoch (i.e. once nonce has been rotated)
- Reproducible state transitions based on "simultaneous" resolution of player actions per epoch.
- "Trailing Verification" of game events, necessitating support for rollback netcode.
- Real-time support can't be faster than the time it takes to aggregate and process the packets sent in a single epoch. EVE's 0.5tps is an example of this in action
- Could support "serverless" architecture of most epoch-to-epoch events in a sort of "cluster" configuration of clients in an area, with the game server acting as an orchestrator/matchmaker.
- ex: players in one of EVE's grids = cluster of users to p2p network and communally agree on gamestate
- 51% attack? How does the server's verification play into this?
- Mitigations for a malicious server/host?
- Unfair/exploitative world state (more/less NPC spawns than normal in an area?) detection
- Diminishing weights on consensus from similar viewpoints?
- ex: 5th member of the party's trailing verification worth 0.001% of the 1st member of the opponent's party
- Servers staked in traditional proof-of-stake model, slashed on consensus verification of foul play
- 3rd party servers acting like "judges" in a card tournament, all sides hand off their version of the validation of the gamestate, judges reproduce based on the initial conditions, award whoever matches their validation.

## Player Data
- Persistent storage of player data in a "universe-wide" blockchain
- Player data as NFT wallet, with handful of "static assets" (EXP, Money, etc) and unbounded list of "variable assets" (e.g. pointers to other NFT instances).
- All kinds of potential for Smart Contracts on assets
- Guilds/Corps/Parties/etc can all be defined via smart contracts, even including loot rules
- slowlorus attacks via smart contracts?
- gas for contract execution = IGC or RWC?

## Cross-server Play
- "Closed"/private servers unable to export unless available to public to enter
- Weird clash with sovreignity & blockchain server staking: does ownership of an ingame territory flipping to another group constitute moving it to another server?
- On sov flip, new owners could associate new "hosting server" for the territory
- Upper limit of supported territories based on stake?
- Server owner conflict of interest? How do we handle someone turning out the lights if they're losing?

## Administration
* Voting token layer for handling network-wide disputes
* server hosts rewarded for length of hosting?
* Epoch-based cash-out, rewarded for not cashing out in voting token?
* Flagging users names/interactions could be handled in tribunal manner, but ideally would remain with the developers/company for final say to retain liability protection.
* Mostly hands-off moderation, server owners given chat moderation authority without need for dev intervention
* Voting token could be used to drive feature development
* Voting tokens would be non-transferrable, only revokeable by vote/company dictate