Choosing the Right Tool - A Comparative Look at gRPC, REST an GraphQL

In the ever-evolving landscape of web development, the choice of technology can drastically impact the design and efficiency of the application. This talk aims to demystify and compare three key players in API communication: gRPC, REST, and GraphQL. We delve into what makes each of these technologies unique, and more importantly, how to determine which is the best fit for your specific project needs.

Transcript

2. Your guiding light to innovation
3. CHOOSING THE RIGHT TOOL: A COMPARATIVE LOOK AT GRPC, REST, AND GRAPHQL JAVA MEETUP - VIENNA DAMJAN GJUROVSKI 21 May 2024
4. 4 HI • Damjan Gjurovski • Dev and Ops J • Written REST code for a long while, dabbled in GraphQL for some frontend apps and implemented (and debugged) gRPC in performance-critical applications
5. WHAT’S IN IT FOR ME 1. The sandbox 2. REST, GraphQL, gRPC – what do the letters mean 3. Let’s compare them, but fairly 4.What should I use in my project?
6. THE SANDBOX 1
7. 7 OUR TINY APP • Book rooms in a hotel • Mental model • What would a database model look like • What would Java entities look like
8. REST, GRAPHQL, GRPC – WHAT DO THE LETTERS MEAN 2
9. 9 REST • Representational state transfer • Resources on the server • Identifiers • Representation over the wire • HTTP verbs to modify state
10. 10 REST PROBLEMS • Misuse / verbs in URL • Batch transactions • Partial data requests
11. 11 GRAPHQL • Only request data as needed • Schema is defined separately, and compatibility can be ensured/enforced
12. 12 GRAPHQL PROBLEMS • Caching • Complex queries • Schema evolution
13. 13 GRPC • Binary protocol – super fast • Protobuf schema is defined separately, and compatibility can be ensured/enforced • Nice features for scale
14. 14 GRPC PROBLEMS • Remote procedure calls • Server to server communication • Code generation / stubs • Debugging
15. LET'S COMPARE THEM, BUT FAIRLY 3
16. 16 • Client – Server • Browser – Server • API Client - Server • Server – Server • Microservices • Backend servers COMMUNICATION TYPES
17. 17 • Lots of data • Only partially used data • Streaming data • Little data • Data represents state DATA
18. 18 • Security • Latency • Schema evolution NON-FUNCTIONAL REQUIREMENTS
19. 19 HOW TO BUILD FAST ON TOP OF SLOW SYSTEMS
20. 20 HOW TO BUILD FAST ON TOP OF SLOW SYSTEMS
21. WHAT SHOULD I USE IN MY PROJECT 4
22. 22 THIS IS THE WAY • Use gRPC for critical, complex projects • Use graphQL when you need to support multiple clients with different needs • When not sure use REST
23. 23 WANT TO KEEP THE DISCUSSION GOING? MESSAGE ME ON LINKEDIN!