Event based semantics:
Event ontology focuses on describing dynamic subject areas and modelling activities: all facts generated by the recognized actors and recorded in the system are presented as a sequence of events forming a directed acyclic graph.
The structure of the graph is determined by the models of entities, properties and actions for which events are generated and validated (no event can be attached to the graph from the outside of the model). The unification of basic models open doors for the creation and support of a single semantic space necessary to unify data exchange.
Conditional communication of events is used to create executable models that generate events only at predefined values of previous events. Business processes and digital agreements are designed based on executable models.
Architecture:
Execution of event models is based on the ideology of dynamic and reactive programming paradigm using conditionally causality of events in the column. Interaction of models is implemented on the principles of Event-driven using a directed acyclic graph as a queue of events. Semantic certainty of events in the column allows for parallel asynchronous interaction of a set of semantically independent (with no conditional links) executable models.
Acceptance of an event as a basic data element automatically provides all the benefits of the Event Sourcing approach: tracking the state by events, storage temporality, and the ability to restore states. Column navigation, semantic search and execution of semantic models are implemented in a special semantic browser.
Peer-to-peer network:
The graph is supported by a peer-to-peer network using user devices as nodes (including smartphones, IoT sensors and data centres). The nodes store the branches of the graph created by the models to which the user subscribes. The division of the event graph into semantic clusters (groups of models belonging to different subject areas) ensures a natural horizontal scaling of the network.
Network transactions are generated and validated according to the semantic models, which enables parallel processing of transactions built on non-interoperable models, as well as data exchange between independent decentralised applications. Network clustering and transaction generation based on semantic models allows the application of customisable consensus algorithms for different types of events.
Cryptography:
Network nodes are identified by a cryptographic key pair: the private key is stored on the node and the public key is stored in a decentralised public key trust network.
Events, when added to the graph, are signed with the user's private key and due to mandatory conditional references to previous events are cryptographically linked in a hash chain, which allows unambiguous identification of change actors and makes data falsification extremely difficult. If necessary, if specified in a semantic model, data encryption is used in a column with role-based access control.
Utility token:
Payment for network resources and the organisation of various exchange processes is implemented using the native network token. Token transactions are confidential: a special cryptographic algorithm provides validation of encrypted amounts in transactions, and there is a mechanism for hiding the recipient's address.
IA and internet of things:
Storing data in a unified event format makes it much easier to use machine learning methods to optimise and analyse data (recognising synonyms or spam), to extract hidden knowledge (identifying new concepts and relationships) and to forecast aggregate data from subject clusters.
Aggregation of data in the branches of the semantic column, produced by actors in the course of their activities, is a unique material for training intelligent agents in the relevant subject areas.
Semantic certainty of network transactions, data protection against falsification, unique identification of actors (including sensors) and objects, and high speed of transaction confirmation (using models that are not legally relevant) can make the event semantics to be basic for the Internet of things.