Building Fastly's Platform for Scale: A Journey of Continuous and Incremental Engineering by Inés Sombra and Patrick Hamann
This talk explores Fastly's journey in building and scaling its platform, highlighting key architectural principles and addressing the inherent challenges of achieving scalable growth. The focus is on understanding how Fastly prioritizes availability, horizontal scaling, data ownership, and a platform-centric approach.
We’ll discuss the critical role of real-time monitoring and user feedback in our engineering cycles, ensuring that our platform evolves in response to actual usage patterns. Through concrete case studies, we’ll illustrate how these practices have led to measurable improvements in performance and user experience.
Join us to learn how Fastly’s dedication to continuous improvement helps create a better internet where all experiences are fast, safe and engaging
Talk Questions
Questions moderation
All questions have to comply with our Code of conduct. So if you don't see your question right after sending it it's because either it has not been moderated yet or it's a question that does not comply with our CoC.
Question 662
What is POP abreviation you use in many places?Question 664
How do you do a load-balancing on a pop?Question 665
Which tool are you using to integrate the microfrontend at the infrastructure level?Question 667
How do you handle the frontend design of each component to maintain an homogeneous look that will not break the cohesion?Question 666
How do you handle dependencies in the micro front-end architecture when you have a large scale?Question 677
Considering AWS’s recent extension of the API Gateway timeout limit beyond 29 seconds(https://aws.amazon.com/about-aws/whats-new/2024/06/amazon-api-gateway-integration-timeout-limit-29-seconds ), do you think AWS Lambda is now suitable for integrating with large language models (LLMs) in real-time applications, such as chat systems? Or would another architecture be more appropriate for handling the computational demands and response times of AI models in production environments?Question 668
Really appreciate the talk. It sounds like you can iterate pretty fast. I am curious to hear your comments on what timescales we are talking about with these lessons learned etc