Explore the implementation of Antimony, a real-time distributed stream processing system written entirely in Rust, in this RustFest Zürich 2017 conference talk. Delve into the architecture and efficiency improvements over Apache Storm, drawing inspiration from Twitter's Heron paper. Learn how Antimony leverages Tokio.rs to create a cross-platform, high-performance, and lightweight solution for stream processing and distributed computation. Discover the motivation behind Antimony's development, its key components such as the Topology Master and Stream Manager, and how it addresses the architectural challenges faced by Storm. Gain insights into defining topologies and using the antimony-cli tool. Presented by Mohammed Makhlouf, an experienced Software and Security Engineer with expertise in machine learning, security, and large-scale distributed systems.
Antimony - A Real Time Distributed Stream Processing System
Overview
Syllabus
Intro
The plan of this talk
The use case for a real time streaming
Attackers / Defenders best friend
Apache Storm
A Streaming Job
Some Terminology
Example topology
Physical Plan of topology
Motivation for Antimony
Some problems with that architecture
Inside the Storm Worker
Zookeeper Overload
Some homegrown wisdom
Two things happened
Antimony Architecture
Topology Master
Stream Manager
Define a topology
antimony-cli
The Name
Thank You
Taught by
Rust
