Resource Management Ecosystem and Cloud Foundry - A Review of the State of the Art
summary
Cloud Foundry treats the application as the unit of deployment: CF push = app abstraction. Behind the scenes containers use cgroups and the kernel fair scheduler, with CPU proportional to memory by default. Diego implements an on‑the‑fly auction and linear optimization (real Go code on GitHub) to pick the best host; Diego is a monolithic scheduler but can integrate with Mesos as a framework. Diego is available in beta.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 0:16The Resource Management Problem
- 1:26Evolution of Resources
- 2:35Changing Demand Models
- 3:35History of Schedulers
- 5:29The Rise of Virtualization
- 6:33VMware Resource Management
- 8:00Enterprise Challenges
- 11:06The Google Influence
- 14:01Scheduler Taxonomies
- 17:22Cloud Ecosystems
- 19:33Cloud Foundry Philosophy
- 22:05CF Resource Management
- 24:35Diego Scheduling Optimization
- 26:12Recommendations for Operators
- 29:06Enterprise Big Data Strategy
- 30:01Future of Resource Management
- 34:35Q&A and Closing
related talks
37mCloud Foundry Diego Overview
01mIntroduction: Cloud Foundry
24mCloud Foundry and Containers (or: Where does Cloud Foundry Fit in the Container Ecosystem?)
32mCloud Foundry Diego Update
15mThe Cloud Foundry Renaissance
22mEnabling Cloud Foundry to Run on Bare Metal
28mDiego vs Eirini: How to Choose the Best Fitting Container...
48mCloud Foundry - your PaaS on OpenStack