JSON is Great, but Protobufs Make It Faster

on YouTube ↗16m

" summary

Shows how Protocol Buffers replace JSON by compiling a .proto schema into language-specific bindings (C++, Java) and sending a binary payload over the network, with concrete numbers: 3x faster serialization (JSON 91 ms vs Protobuf 30 ms), 13x faster deserialization (JSON 50 ms vs Protobuf 4 ms), and 40% smaller payloads (JSON 9000 B vs Protobuf 5500 B). Explains required/optional fields, backward compatibility, and that real microservices (LinkedIn example) see large latency reductions.

AI-written summary · about

" chapters (9)

1. 0:00Introduction and speaker intro
2. 0:51Why JSON has scaling issues
3. 3:10Introducing Protocol Buffers
4. 5:15How Protocol Buffers work
5. 7:41Defining Protobuf schemas
6. 9:27LinkedIn’s Protobuf success case
7. 10:56Performance comparison numbers
8. 13:07When to use Protobufs
9. 15:54Q&A session

" related talks

• 35mBeyond JSON: Improving inter-app communicationAaron Quint · Full Stack Fest 2015
• 29mBinary Protocols - The Current State of AffairsJason Smith · CloudNativeCon + KubeCon Europe 2017
• 11m"RPC Frameworks Overview"Yulia Oletskaya · RubyConfBY 2019
• 38mA Quick Guide to RPC FrameworksYulia Oletskaya · Birmingham on Rails 2020
• 46mIntroduction to gRPCMete Atamel · Codemotion Amsterdam 2017
• 49mBuilding fast microservices with binary protocol - introduction to gRPCRay Tsang · Oredev 2017
• 46mA Different Kind of Networking with gRPC and Protocol Bufferskmats · DroidKaigi 2018
• 34mRuby and gRPCJoy Paas · RubyConf Philippines 2018