The Maximal Graph - Wilker Silva

Learn how Nubank's Maximal Graph system handles complex data relationships and queries, utilizing a triplet store, EQL, resolvers, and GraphQL services for scalability and decentralization.

Key takeaways

GraphQL is a custom language with a typed schema and supports introspection, enabling features like auto-complete and cool introspection features.
The Maximal Graph is a system created by Nubank to handle the complexities of large, distributed systems and rich demands.
The system uses a triplet store and EQL (Entity Query Language) to define properties and relationships between entities.
Properties are represented as entities themselves, with their own meta-levels of description.
The system uses resolvers, which are edges connecting properties, to define how data is fetched and transformed.
GraphQL services can be used to handle complex queries and provide a way to batch requests.
The system uses a query root mapping to access GraphQL from the root, and has support for mutations.
The Maximal Graph is a feature-rich system that can handle complex data relationships and queries.
RDF (Resource Description Framework) and RDFS (Resource Description Framework Schema) were discussed as other systems that can handle similar complex data relationships.
The system has a strong focus on scalability and handling large amounts of data.
The speaker emphasized the importance of using a typed schema and supporting introspection in GraphQL.
The system uses a property-centric approach, where properties are treated as entities themselves, rather than as attributes of an entity.
The speaker discussed the importance of avoiding collisions between property names, and using prefixes and aliases to resolve this issue.
The system has a strong focus on decentralization and allowing multiple services to interact with each other.

The Maximal Graph - Wilker Silva

More talks