BestHub
Sign in
Volcano Engine Developer Services
Volcano Engine Developer Services
Oct 12, 2023 · Cloud Native

How ByteDance’s Katalyst Memory Advisor Boosts Kubernetes Memory Efficiency

This article explains the challenges of memory management in mixed workloads, outlines the limitations of native Linux and Kubernetes mechanisms, and details how ByteDance’s open‑source Katalyst Memory Advisor improves memory utilization, QoS, and eviction handling through user‑space policies, multi‑dimensional interference detection, and adaptive mitigation actions.

KatalystKubernetesResource Optimization
0 likes · 17 min read
How ByteDance’s Katalyst Memory Advisor Boosts Kubernetes Memory Efficiency
ByteDance Cloud Native
ByteDance Cloud Native
Oct 11, 2023 · Cloud Native

How Katalyst Memory Advisor Optimizes Kubernetes Memory Management in Mixed Workloads

This article explains the challenges of memory management in mixed Kubernetes workloads, introduces ByteDance's open‑source Katalyst Memory Advisor, details native allocation and reclamation mechanisms, outlines its architecture and plugins, and describes interference detection and multi‑level mitigation strategies to improve memory utilization and service quality.

KatalystKubernetesResource Optimization
0 likes · 19 min read
How Katalyst Memory Advisor Optimizes Kubernetes Memory Management in Mixed Workloads
ByteDance Cloud Native
ByteDance Cloud Native
Aug 15, 2023 · Cloud Native

What’s New in Katalyst v0.3.0? Core Enhancements Explained

Katalyst v0.3.0 introduces major upgrades including enhanced KCNR API bandwidth isolation, a more extensible task and async execution framework, advanced mixed‑deployment controls, load‑aware resource prediction, and concurrent unit testing, all aimed at improving cloud‑native resource management efficiency.

KatalystKubernetesresource management
0 likes · 4 min read
What’s New in Katalyst v0.3.0? Core Enhancements Explained