GraphQL และ gRPC

RESTful APIs are widely used, but they aren't always the perfect fit for every scenario.

Sometimes, you need more control over data fetching or require extremely high performance for inter-service communication.

This lesson explores two powerful alternatives: GraphQL and gRPC.

Traditional REST APIs often lead to two common issues:

• Over-fetching: Receiving more data than needed in a single request, wasting bandwidth.
• Under-fetching: Needing to make multiple requests to different endpoints to gather all required data.

This can be inefficient, especially for mobile clients or complex UIs.

GraphQL is a query language for your API, and a server-side runtime for executing those queries.

It empowers clients to request precisely the data they need, eliminating over-fetching and under-fetching.

Think of it as filling out a form for your data request, rather than accepting a pre-defined bundle.

Every GraphQL API is built around a schema. This schema defines all the types of data and operations (queries, mutations) available.

It acts as a contract between the client and the server, ensuring clients know exactly what they can ask for.

Here's a simple example:

Clients use GraphQL queries to fetch data. They specify the exact fields they want from the available types defined in the schema.

This means a single query can replace multiple REST requests.

See how we only ask for name, email, and post titles:

Just as queries fetch data, mutations are used to modify data on the server. This includes creating, updating, or deleting records.

Mutations also return data, often the newly created or updated object, confirming the change.

Here's an example to create a new post:

gRPC (Google Remote Procedure Call) is an open-source framework for high-performance communication between services.

Unlike GraphQL, gRPC is more commonly used for internal microservice communication where speed, efficiency, and strict contracts are paramount.

It's built on HTTP/2 for transport and Protocol Buffers for message serialization.

gRPC uses Protocol Buffers (Protobuf) as its Interface Definition Language (IDL) and its underlying message interchange format.

Protobuf defines the structure of your data and the service interfaces in a language-agnostic way.

This definition is then used to generate client and server code in various languages.

Key features of gRPC for high performance:

• HTTP/2: Enables multiplexing (multiple requests over one connection) and bidirectional streaming.
• Binary Serialization: Protobuf serializes data into a compact binary format, which is faster and smaller than JSON.
• Code Generation: Automatically generates client and server stub code from .proto definitions, simplifying development.

When to use which?

• GraphQL: Ideal for flexible, client-facing APIs where clients need to define their data requirements. Excellent for mobile apps and complex UIs.
• gRPC: Best for high-performance, low-latency inter-service communication (e.g., microservices), or when strict contracts and efficient data transfer are critical.

They solve different problems and can even be used together in a larger system!

Consider a scenario where you are building a new mobile application with a complex UI that needs to fetch varied data from a backend with minimal network requests. Which API style would be most suitable for the client-server communication?

In this lesson, we explored two powerful API alternatives: GraphQL and gRPC.

• GraphQL provides a flexible query language, allowing clients to fetch precisely the data they need, tackling over- and under-fetching issues.
• gRPC offers a high-performance, binary communication framework, ideal for efficient inter-service communication using HTTP/2 and Protocol Buffers.

Understanding their strengths helps you choose the right tool for different system design challenges.