Overview:
Orbital is a data gateway that automates the orchestration, integration, and transformation of data and services across an enterprise. It is built to remove glue code by using existing API specifications enriched with semantic metadata. Instead of defining centralized mapping code, Orbital creates integrations on the fly. It is designed for teams that work with multiple API protocols—REST, gRPC, SOAP, Kafka—and want an integration layer that updates automatically as service specifications evolve. It does not require shifting an entire tech stack to GraphQL.
Core Features:
Spec-driven integration: Orbital powers integration flows based on existing API specs (OpenAPI, Protobuf, Avro, JSON Schema) enriched with semantic metadata created in Taxi.
Decentralized orchestration: There is no central mapping code; integration logic is derived from metadata embedded in each service's API spec, allowing teams to ship changes independently.
Automatic adaptation: Integration flows adjust automatically as API specs change, so consumers remain unaffected by backend updates.
Spec publication options: API specs can be published to Orbital via polling, Git repository integration, or direct push from an application.
On-the-fly query composition: Orbital builds integration for each query and composes APIs on demand, with no resolvers to maintain.
Technology-agnostic support: Works with REST, gRPC, SOAP, Kafka, and databases without requiring a unified protocol or migration.
Use Cases:
Federating data sources for teams: Developers can query a single API that pulls from multiple underlying services (databases, message brokers, APIs) without writing glue code.
Automating integration between evolving microservices: Teams that change their APIs over time can have integration flows update automatically, reducing breakage.
Replacing a central integration team: Decentralized teams maintain their own API specs with semantic metadata, eliminating the need for a dedicated middleware team to manage resolvers.
Why It Matters:
Orbital addresses the brittleness of glue code by automating integration directly from API specifications. It does not require migrating to a new protocol like GraphQL and avoids the need for a separate shim layer to adapt existing services. Because the integration logic is derived from decentralized team-owned specs, there is no single mapping codebase to maintain. This makes it a practical option for organizations that want a self-adapting data federation layer built on their existing technology stack.
