May 11, 2022

Exploring the Underlying Network of Phoenix Applications

clicks

Source: youtube.com

Mat Trudel dives into the intricacies beneath Phoenix applications, exploring BEAM's OTP patterns and supervisory structures that govern network interactions. Starting with an introduction to his Elixir-based web server, Bandit, he expands on how it interfaces with Phoenix through the Plug API. Trudel discusses Bandit's functionalities, comparing it to the Cowboy server and detailing the integration process with Phoenix. The talk encapsulates the essence of Phoenix as an HTTP-centric system where requests are isolated and managed through separate BEAM processes. Trudel elaborates on the architectural layers of web servers, distinguishing between protocol-specific tasks and generic connection handling while underscoring the scalability and robustness of the Elixir/Erlang ecosystem. He elucidates on the Thousand Islands library, which supports socket management, and how it empowers Bandit to run efficiently. Wrapping up, Trudel encourages developers to demystify the network stack, advocating for the comprehensible and approachable nature of these technologies.

Exploring High-Dimensionality in Language Design

Brooklyn Zelenka discusses innovative approaches to language design, advocating for the creation of small, composable DSLs and structured languages to simplify coding complexity.

May 11, 2022

Understanding Networking on the BEAM with Elixir

Andrea Leopardi presents insights into how the BEAM's architecture is particularly suited for network applications. He discusses the handling of TCP connections and the design patterns for building scalable and fault-tolerant systems.

Jul 10, 2023

Exploring Machine Learning in Elixir with a Recommendation Engine

Andrew Forward discusses the potential for using Elixir to build a recommendation engine for a gifting platform. He explores traditional machine learning algorithms such as KNearestNeighbour, (Naive) Bayes, and KMeans, and introduces the Elixir library 'scholar' for implementing these algorithms.

Jul 10, 2023