Moving to Microservices: Optimizing Network Calls

on YouTube ↗33m

" summary

Discusses the network costs and new failure modes that appear when a PHP monolith coexists with many microservices, and practical optimizations to reduce network overhead. Covers how PHP’s curl and curl_multi_exec work, use of the Guzzle wrapper and Avito’s usage, choosing correct timeouts and their failure effects (worker-pool exhaustion, cascading failures), and using a reverse proxy with PHP for keep-alive, DNS-resolve savings, and retries.

AI-written summary · about

" chapters (12)

1. 0:00Introduction
2. 1:43Why network optimization matters
3. 4:57Sequential requests problem
4. 7:04Parallel requests with Go and batch
5. 10:18Curly and optimizations
6. 12:45Reverse proxy setup
7. 15:15Graceful degradation approach
8. 18:27Retry strategies and risks
9. 22:44Retry implementation mistakes
10. 25:34Microservices load impact
11. 27:43File strategy for microservices
12. 28:47Summary

" related talks

• 48mMicroservices vs The Distributed MonolithChris Riley · Dutch PHP Conference 2018
• 34mMigrating a High‑Load Project to MicroservicesDmitriy Nemesh · PHP fwdays'17
• 43mGo micro with MicroservicesMariusz Gil · International PHP Conference Spring 2017
• 36mRemoving the legacy, the microservice approachNiels van Esch · Dutch PHP Conference 2017
• 46mFrom a PHP Monolith to Scala MicroservicesDenis Ivanov · HighLoad++ 2015
• 37mHow to run 200+ PHP services in production and not lose your mindYurii Panaiotov · PHP fwdays'25
• 19mMicroservices vs. Monolith for Scalability and EfficiencyAsif Mehboob · Conf42 Machine Learning 2025
• 49mFrom Microservices to MigroservicesFrançois Teychene · DevFest Lille 2019