Back to catalog
Orbital
Orbital is an open-source data gateway alternative to Apollo GraphQL that connects APIs and databases on the fly using semantic Taxi specifications.
Open-source alternative to — compare it against:
Overview
Apollo GraphQL demands bespoke resolvers and a unified schema layer to stitch backend services together. Orbital infers data relationships on the fly using semantic tags embedded directly in existing API specifications. This shift turns multi-system data queries into automatic operations without requiring engineers to rewrite application services.
Why Orbital instead of Apollo GraphQL
Under the performance lens, the fundamental difference lies in execution path planning and query dispatch. Apollo GraphQL relies on pre-compiled schema graphs where routers dissect incoming GraphQL queries into static execution plans across subgraphs. When downstream services return nested records, engineering teams must maintain custom batching mechanisms like DataLoader to prevent N+1 query storms and network exhaustion. Orbital avoids handwritten resolution paths by generating dynamic query plans through TaxiQL. When a query arrives, Orbital’s engine determines the shortest resolution graph across SQL databases, Kafka streams, and REST or gRPC endpoints, firing concurrent upstream calls and joining payloads in flight.
This architectural difference creates clear throughput and latency tradeoffs. Apollo GraphQL delivers lower per-query CPU overhead when answering predictable front-end requests against dedicated GraphQL subgraphs, especially when paired with edge-cached query hashes and normalized client caches. Orbital shifts that burden to its runtime planner: calculating query graphs dynamically introduces engine overhead on every ad-hoc request, but eliminates intermediate network hops and middleware services when querying mixed data stores. The performance ceiling in Orbital is determined almost entirely by the slowest downstream datasource rather than gateway middleware, though unindexed relational queries or slow third-party endpoints will directly throttle end-to-end response times.
What it does well
- Automated data joins without custom glue code, allowing non-technical analysts to access unified records without waiting for engineering backlogs.
- Living data catalog generated from existing OpenAPI and database schemas, providing non-technical teams with an up-to-date directory of available business data.
- Spec-driven query generation that updates automatically when underlying APIs change, ensuring business dashboards and reports do not break during routine service updates.