I've started publishing some documentation on the
wiki, along with making public information on our data storage solution. Our data platform is called PlumeDB. This is a layered solution that rides on top of the P2P message bus in the core (e.g., the existing Bitcoin-based messaging system).
http://wiki.dfx.io/display/XAI/PlumeDBWe layer "TreeSpaces" (a multi-dimensional trie index, or prefix-hash tree on multiple attributes) on top of a Kademlia-based DHT implementation sans UDP messaging (funneled over existing P2P core).
Plume supports three core data structures:
- JSON data (e.g., similar to MongoDB)
- Time series data (e.g. realtime data, sensor data, market data, similar to Druid.io)
- Atoms (e.g. OpenCog AtomSpaces data)
Slave Nodes host the treespaces and earn the network fees (Data Reservation Fees). The data originator can choose how many slave nodes are needed to support their use case (default is eight). Private data reservation fees are in XAI per kilobyte-hour paid to each slave node in the data plume. Public data reservation fees charged to the originator are in XAI per kilobyte paid to each slave node in the data plume, and public data access fees charged to consumers are in XAI per query-hour.
Fee Type | Default Rate |
Private Data Reservation Fee | 0.00001 XAI per kilobyte-hour |
Public Data Reservation Fee | 0.0000001 XAI per kilobyte |
Public Data Access Fee | 0.001 XAI per query-hour |
The next major feature release with this functionality will be dropped on testnet in the near future. XAI coins from the new testnet will be distributed to crowdfunders on the Slack group. They will then offer these as invites in the coming weeks if you're not in the slack and want to have some coins to get access and start playing with the features on testnet