Mostly for Jason's brain.


- Control Node

  1. Control Structure
  2. Gateways talk to control node groups.
  3. Control node preference selected by user.
  4. Database management
  5. Databases pulled by control node.
  6. Gateways report status to control node.
  7. Control nodes prefer cloud access, go into offline when cloud access down.
    1. Control node uses downloaded domain database when offline.
    2. All entries have timestamps, and timestamps are used for conflict resolution.

- User management

  1. Users on cloud site.
  2. Customers have domain names (verified)
    1. If not, then customers are assigned a generated subdomain.
    2. Customers may have multiple domains.
    3. Users are unique to domain entries.
    4. It goes on from here, increasing levels of complexity and so on.
  3. Domains contain control groups and users and domain-wide settings.
    1. Control group settings override domains.
    2. Gateways override control groups.
    3. Groups override gateways.
    4. Users override groups.
    5. Devices override users.

- Message passing.

  1. Peer to peer message passing infrastructure.
  2. Only control group master talks to cloud.
  3. All others talk to control group master.
  4. Gateways can also relay, maybe using p2p msg passing daemon.

- Throttling

  1. Burst/stream rates.
  2. Auto-calculated oversubscribing.
scratch.txt · Last modified: 2013/07/20 05:32 by jason
Recent changes RSS feed Debian Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki