3 API Protocols You MUST Know for System Design
43sQuick, high-value tip for interview prep that viewers will want to save and share.
▶ Play ClipAI Summary
The video explains the three main API protocols—REST, GraphQL, and gRPC—and their appropriate use cases in system design interviews. It emphasizes that each protocol has a specific job and choosing the right one is crucial for interview success.
Chapters
[00:02]
REST for CRUD
REST is the default choice for CRUD operations, using resources as URLs and verbs as actions. It rides on HTTP, inheriting caching, retries, and status codes.
[00:15]
GraphQL for flexible data
GraphQL is ideal for product UIs needing flexible, nested data. Clients query exactly the shape of data they want, avoiding over-fetching or under-fetching common with REST.
[00:28]
gRPC for internal services
gRPC is for internal service-to-service calls requiring speed and strict contracts. It uses binary over HTTP/2, generates typed clients, and supports streaming and multiplexing.
[00:43]
Performance comparison
gRPC is significantly faster than JSON over REST for millions of calls between services, making it the choice for high-throughput inter-service communication.
[00:55]
Interview guidance
In system design interviews, choose REST for CRUD, GraphQL for flexible client data needs, and gRPC for inter-service communication at scale.
Understanding when to use REST, GraphQL, and gRPC is essential for system design interviews. Each protocol serves a distinct purpose, and selecting the right one demonstrates architectural thinking.
Mentioned in this Video
Study Flashcards (4)
What is REST best suited for?
easy
Click to reveal answer
What is REST best suited for?
easy
Click to reveal answer
CRUD operations, using resources as URLs and verbs as actions.
00:02
What problem does GraphQL solve?
medium
Click to reveal answer
What problem does GraphQL solve?
medium
Click to reveal answer
It allows clients to query exactly the shape of data they want, avoiding over-fetching or under-fetching.
00:15
What is gRPC designed for?
medium
Click to reveal answer
What is gRPC designed for?
medium
Click to reveal answer
Internal service-to-service calls requiring speed and strict contracts.
00:28
Why is gRPC faster than REST?
hard
Click to reveal answer
Why is gRPC faster than REST?
hard
Click to reveal answer
It uses binary over HTTP/2 and supports streaming and multiplexing.
00:43
💡 Key Takeaways
⚖️
REST as default
Establishes REST as the baseline for CRUD, a fundamental principle in API design.
00:02
💡
GraphQL flexibility
Highlights the key advantage of GraphQL in reducing over-fetching, a common pain point.
00:15
📊
gRPC for scale
Emphasizes the performance benefits of gRPC for high-volume inter-service communication.
00:28Full Transcript
[00:02] your system design interview, you might be in trouble. There's actually three protocols that are worth knowing, and each one has a job. REST is certainly your default choice for CRUD operations. Resources as URLs, verbs as actions, and
[00:15] it rides on top of HTTP, so you inherit caching, retries, and status codes for free. But then there's GraphQL, which is great for product UIs that need flexible, nested data. Clients can write a query describing exactly the shape of
[00:28] the data that they want, and the server returns that, no more or no less. So you like there might be with REST. Lastly, there's gRPC. It's for internal service-to-service calls where you want speed and strict contracts. Binary over
[00:43] HTTP/2, protocol schemas generate typed clients in every language, and you get streaming and multiplexing built in. It's significantly faster than JSON over REST when you're making millions of calls between services. So in your
[00:55] for GraphQL when the client needs flexible or deeply nested data, and drop in gRPC when you're designing inter-service communication at scale. either over on YouTube or on hellointerview.com.
