Caicloud容器云系统的部署实践.pptxVIP

构建基于Kubernetes的容器云系统 Containers in the Spotlight Hardware Disk Linux NIC vHardware vDisk Baremetal NIC vNIC Disk Application Send() Read() Reboot() Hypervisor Application … Hardware Virtual Machine Reboot() Linux Send() Read() Hardware NIC Child Mnt NS Child Mnt NS Child Net NS Child Net NS Capabilities / Seccomp Containers in the Spotlight Send() Read() Reboot() Bins/Libs Routing Process Disk Container: Runtime + Image Using containers for a decade Running 2 billion containers a week Solves application migration nightmare Saves billion dollars a year Docker: 5x yearly growth rate Case Study: Containers in Google Solved Problems of the World? Case Study: Cluster Management in Google 1 SRE handles ~ 10,000 machines with 99.999% reliability Clustering is the hard part. In Google: NO team dedicated for container study HUNDREDS of engineers built THREE cluster manager systems HUNDREDS of teams building ecosystems Clustering is the real value in building serious production container systems All the Fun Stuff Began Here Kubernetes Design Principles declarative imperative simple complex labels hierarchy legacy compatible extensible and pluggable application centric Open Source Cluster Management Projects out There Things Swarm, Mesos and Kubernetes do in common (more or less) • Service discovery • Scheduling and orchestration • Scaling • Cross-host networking and communication However, • Swarm is Docker native BUT hard to extend • what if I want to use rkt? vendor lockin! • Mesos is not designed for containers • for resource sharing More concretely, Kubernetes empowers more: Config Containers App containers Volumes Kubernetes Highlights How to group resources? • Pods Lessen learned from Borg • Jobs are usually grouped • e.g. log offloading • Allow teams to develop distinct part of application • improve robustness • e.g. log can be offloaded even if container crashes • At