Mesos: The Operating System for your Cluster
summary
Mesos claims to act like an OS for a data center: tens of thousands of CPUs cluster-scale deployments, Docker-based packaging, and an offer model that inverts requests so frameworks receive resource offers. It implements two-level scheduling (Mesos allocates resources, frameworks decide tasks), supports Marathon, Chronos and Aurora (Aurora running most Twitter services), and is used in production at large scale; persistent offers and preemption are under development.
AI-written summary · about
chapters (22)
- 0:00Start
- 1:18A long time ago...
- 3:22Mesos: an Operating System
- 3:41Isolation
- 5:14Resource Sharing
- 6:27Common Infrastructure
- 8:21Distributed System* Anatomy
- 9:03Static Partitioning
- 9:42Mesos: a Level of Indirection
- 11:37Tons of Flexibility!
- 11:58Jenkins
- 13:01Hadoop
- 14:39Marathon
- 16:06Chronos
- 17:26Aurora
- 18:50Spark
- 20:13How do I use these?
- 22:06Problems with Specifications
- 27:38Language Bindings
- 28:07Should I build it on Mesos?
- 28:43But should I really build it on Mesos?
- 29:43The Plan, redux
related talks
59mBuilding Distributed Systems with Apache Mesos
45mMesos, Then and Now
53mDocker as the Building Block for Datacenter-Scale Applications
01h 00mDeveloping Distributed Systems with Apache Mesos
57mA Distributed and Fault-Tolerant Application with Apache Mesos & Apache Aurora
39mHow we learned to love the Data Center Operating System
22mKeynote: The Future of Apache Mesos and DC/OS
19mKeynote: Operating 600+ Mesos Servers on 7 Datacenters @Criteo