-
Notifications
You must be signed in to change notification settings - Fork 2
Use cases
Maxim Orlovsky edited this page Jan 29, 2019
·
4 revisions
Pandora Boxchain will have the following use cases, each of which will have its own workflow:
- Creating and updating data structures
- Creating/registering kernels and datasets
- Updating kernels and dataset details (price updates only?)
- Working nodes workflows
- Cognitive computing workflow
- Cognitive computation arbitration
- Reviews and Plagiarism
- Reviewing kernels and datasets by their users
- Plagiarism arbitration
- Governance
- Registering worker node
- Protocol changes
Each of use cases (shown by ellipses) has its own workflow, described in separate issues. Each of workflows can be seen as a game (in terms of games theory), played by actors (human figures on the UML schemes below).
Thus,
- links between actors and use cases show which players participate in which games
- links between actors show which of them are subtypes of other actor types
- Nodes
- Worker Node — node performing cognitive computing tasks
- Validator Node — worker node with sufficient stake validation procedures
- Arbiter Node — validating node with a high reputation
- Governor Node — arbitrating node with the highest reputation
- Full Node — node with a complete copy of blockchain but that does not participate in cognitive work (and cognitive mining)
- Light Node — a node that does not participate in cognitive work (and cognitive mining) and does not have a complete copy of the blockchain
- Data Providers
- Kernel Providers
- Dataset Providers
- Clients - they create and pay for cognitive work. They can also be simultaneously data providers
- Data entries, including kernels and datasets
- Kernels, consisting of model and weights
- Datasets
- Dataset batches, created by the Clients from the dataset for a specific Cognitive Order
- Cognitive data - output from a cognitive task
- Inputs: kernels and datasets
TBD