Now my question is, what is the best architecture to choose for this database?
Mostly because you don’t have any other options. Unless you get TCP/HTTP the only option for now is in memory db’s. sqllite option could work as well but I’m fairly sure it won’t scale. (eg: augur core).
So a graph database architecture seems like a good choice. But is it the best choice? What other architectures have you considered? What led you to the graph database?
For scaling if @cryptoschindler suggestion can be implemented then we’re gold. Also most of the benchmarks shows that graph databases are way faster. Check :::https://www.researchgate.net/publication/220996559_A_comparison_of_a_graph_database_and_a_relational_database_A_data_provenance_perspective
You mentioned creating an ORM. Personally, I’m more interested in just having the query language or the low-level API.
The ORM in question is not the magic ORM you’re thinking of.
Basically I want the graph structure be a simple structure and the ORM part to be added on top of the graph. For example for our own API we need custom validations per field per action, sanitisation and a lot of custom logic. See Idea for CRUD types - feedback appreciated
The idea is to create the ultimate GraphQL generator.
For a graph database a GraphQL layer is esentially useless. Because you don’t need custom resolvers for complex queries, parsers etc. See Cypher is the best example as why you don’t need it. Connecting your React app to Neo4j with React Hooks | Neo4j Developer Blog
Baby steps.
PS: @cryptoschindler I’ll be in touch over a message once I clean my stuff.