Difference between revisions of "WP/Containers density"
< WP
(Created page with '= Containers density = Why OpenVZ containers density is so much better that that of Virtual Machines (Xen, VMware, KVM). - automatic elastic behaviour and no memory reservation -…') |
|||
| Line 1: | Line 1: | ||
= Containers density = | = Containers density = | ||
| − | + | ||
| − | + | == What is density? == | |
| − | + | Density is a characteristics which tells one how many containers (CTs) or Virtual Machines (VMs) virtualization technology can run successfully on the given hardware. Typically users also care for quality of service of their software, i.e. how well their services are working on utilized node and how fast they respond to external requests. Average response time, 99.9% requests response time and max response time are typical metrics of quality of service. When hardware node is capable to bear the load these metrics either do not grow much with bigger number of containers or grow linearly. When hardware node becomes over utilized these metrics typically start to degrade exponentially (e.g. due to memory swap out). | |
| − | + | ||
| − | + | == What makes containers to be perfectly suitable for high density? == | |
| − | + | # Containers do not reserve memory assigned to it. They exhibit elastic behavior, if applications do not use | |
| − | + | ||
| + | # explain that even best hypervisors (ESX) achieve only 2x memory overcommit. Actually 50% of it are due to page sharing, and 50% due to balloon. | ||
| + | |||
| + | # Virtuozzo templates and file sharing add even more savings | ||
| + | # what happens when CT is above it's limit | ||
| + | # what happens when node RAM is exhausted | ||
| + | # plots and examples. Kir had http_load plot in the past. We will have LAMP results as well. | ||
Revision as of 17:56, 14 March 2011
Containers density
What is density?
Density is a characteristics which tells one how many containers (CTs) or Virtual Machines (VMs) virtualization technology can run successfully on the given hardware. Typically users also care for quality of service of their software, i.e. how well their services are working on utilized node and how fast they respond to external requests. Average response time, 99.9% requests response time and max response time are typical metrics of quality of service. When hardware node is capable to bear the load these metrics either do not grow much with bigger number of containers or grow linearly. When hardware node becomes over utilized these metrics typically start to degrade exponentially (e.g. due to memory swap out).
What makes containers to be perfectly suitable for high density?
- Containers do not reserve memory assigned to it. They exhibit elastic behavior, if applications do not use
- explain that even best hypervisors (ESX) achieve only 2x memory overcommit. Actually 50% of it are due to page sharing, and 50% due to balloon.
- Virtuozzo templates and file sharing add even more savings
- what happens when CT is above it's limit
- what happens when node RAM is exhausted
- plots and examples. Kir had http_load plot in the past. We will have LAMP results as well.
