briefing: containers
TRANSCRIPT
Briefing: Containers
Datacloud, Monaco 2015 David Mytton, CEO, Server Density
David Mytton
Agenda
• What? New technologies
• Where? Cloud and platforms
• When? Use cases
• Why? History
Why?
Why?• Isolation - single process to full OS
Why?• Isolation - single process to full OS
• Resource control
Why?• Isolation - single process to full OS
• Resource control
• Prioritisation
Why?• Isolation - single process to full OS
• Resource control
• Prioritisation
• Accounting
Why?• Isolation - single process to full OS
• Resource control
• Prioritisation
• Accounting
• VMs? Less overhead
History• 15 years
History• 15 years
• 2000: FreeBSD jail
• 2004: Solaris zones
• 2001: Linux VServer
History• 15 years
• 2000: FreeBSD jail
• 2004: Solaris zones
• 2001: Linux VServer
• 2007: cgroups
• 2008: LXC
New technologies• 2013: Docker
New technologies• 2013: Docker
• Packaging / format
• API (libcontainer)
New technologies• 2013: Docker
• Packaging / format
• API (libcontainer)
• 2014: Rocket (rkt)
New technologies• 2013: Docker
• Packaging / format
• API (libcontainer)
• 2014: Rocket (rkt)
• CoreOS, Ubuntu Core, Red Hat Atomic
Platforms• Why? Scheduling
Platforms
• Apache Mesos
• Why? Scheduling
Platforms
• Apache Mesos
• Why? Scheduling
• Kubernetes
Cloud
• Google: Container Engine
• AWS: EC2 Container Service (ECS)
Cloud
• Resources
• Scheduling
• Failover
• Other cloud services
Cloud
“Our customers have told us that scaling and operating these data storage systems is very challenging. Furthermore, they felt that this was undifferentiated heavy lifting and would rather focus their energy on running their applications and growing their businesses.”
- Werner Vogels, CTO, Amazon
Cloud• AWS: EC2 Container Service (ECS)
• Broad AWS adoption
• Less experience with containers
Cloud• Google: Container Engine
• Smaller GCP adoption
• Extensive experience with containers
• Kubernetes ecosystem
When? Use cases• Development == testing == production
When? Use cases• Development == testing == production
• Easier deployment
When? Use cases• Development == testing == production
• Easier deployment
• Easier management (scheduler)
When? Use cases• Development == testing == production
• Easier deployment
• Easier management (scheduler)
Very early!