YoVDO

Highly Available Architectures for Online Serving in Ray

Offered By: Anyscale via YouTube

Tags

Distributed Systems Courses Fault Tolerance Courses High Availability Courses Load Balancing Courses Cluster Computing Courses Ray Serve Courses Anyscale Courses

Course Description

Overview

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore highly available (HA) serving, a key feature introduced in Ray 2.0, in this informative 34-minute talk from Anyscale. Dive into the architecture and functionality of HA serving within single Ray clusters, which eliminates single points of failure and enhances efficiency. Learn how this advancement reduces disruption during head-node failures and improves overall cluster performance. Discover the practical aspects of deploying HA serving in Ray 2.0, and gain insights into its supported functionality. Perfect for developers and system architects looking to optimize their online serving workloads using Ray Serve.

Syllabus

Highly available architectures for online serving in Ray


Taught by

Anyscale

Related Courses

Patterns of ML Models in Production
PyCon US via YouTube Deploying Many Models Efficiently with Ray Serve
Anyscale via YouTube Modernizing DoorDash Model Serving Platform with Ray Serve
Anyscale via YouTube Ray for Large-Scale Time-Series Energy Forecasting to Plan a More Resilient Power Grid
Anyscale via YouTube Enabling Cost-Efficient LLM Serving with Ray Serve
Anyscale via YouTube