Saturday, August 11, 2007

How openQRM helps to save the climate

Everybody is talking about the climate change and how to save the world by reducing pollution and energy-consumption. How openQRM can help ?

I thought about it a while and came to the conclusion that openQRM actively supports reducing the power-consumption by its following key features:

1. Server consolidation

With its automated provisioning technics openQRM takes server consolidation to the next level. By abstracting the data-center services into virtual environments including their hard- and software-profile plus deployment parameters such as number of systems, cpu/memory requirements, high-availability and eventual service-level agreements openQRM assists the system-administrator to find unused and/or under-utilized resources to either power them down or to migrate their services to other servers.

2. Service consolidation

openQRM supports to automatically consolidate critical services in a high-available way. In an active/passive setup openQRM will automatically fail-over its server-services so that services integrated within openQRM via a plugin automatically gain high-availability.

Let's have an example with the following critical services in a data-center :

dhcpd-server for ip-address assignment
tftp-server for network booting
puppet-server for configuration management
nagios-server for enhanced monitoring
linuxcoe-server for automatic installations

Normally a system-administrator would setup 5 dedicated servers plus 5 hot-standbys to provide high-availability in active/passive mode.

==> This means 10 physical servers.

With openQRM you have (at least) 2 options :

I) Consolidating all services on the openQRM-server

1 openQRM-server with the dhcpd-, tftpd-, puppet-,
nagios- and linuxcoe-plugins installed

1 hot-standby for the openQRM-server

In case of a fail-over openQRM will automatically take over the services from all installed/enabled plugins.

==> This would reduce the amount of needed systems from 10 to 2

II) Using the high-availability pool of openQRM

1 openQRM-server
1 hot-standby to have high-availability for the main management system.
5 dedicated systems to deploy the dhcpd, tftpd, puppet, nagios and linuxcoe server
1 hot-standby for all managed servers.

==> This would reduce the amount of needed systems from 10 to 8
with the option to add even more services in high-availability mode without the need to add more hot-standbys (of course it is an option to add more hot-standbys)

3. Migrating to virtual-machines

Another possibility to consolidate services of a data-center is to migrate them from physical systems to virtual machines using various, different virtualization methods. openQRM support all current mainstream virtualization technologies (VMware, Xen, Linux-VServer and QEMU/KVM) and gives the system-administrator a strong tool to move from physical to virtual. openQRM's unique virtualization layer, called "partition engine", is a logical abstraction to physical and virtual resources and conforms the different virtualization technologies within a single management platform. That means the system-administrator can select the virtualization method which fits best for each specific service and may decide at any time to migrate it from p2v, back from v2p and also from one virtualization type to another by just changing the resource-type of a virtual-environment in its profile.

Let's come back to our example. Here one more option using virtualization :

III) Using virtualization for service consolidation

1 openQRM-server
1 hot-standby to have high-availability for the main management system
1 system to deploy a virtualization-host on
1 hot-standby for the virtualization-host

In this setup the system-administrator will create and start 5 virtual machines on the virtualization-host. On each of those partitions of the virtualization-host one of the 5 servers (dhcpd, tftpd, puppet, nagios and linuxcoe) will be deployed. Additional the system-administrator may create additional 1-5 (or more) partitions on the virtualization-host acting as hot-standbys for the 5 critical (virtualied) servers. In case of a error e.g. of the puppet-server the puppet-server will automatically fail-over to another (hot-standby) partition. In case of a failure of the virtualization-host openQRM will automatically fail-over the virtualization-host to its extra (physical) hot-standby.

==> This would reduce the amount of needed systems from 10 to 4
with the option to migrate more services to virtual-machines.

4. Power-management plugins

Last but not least openQRM provides plugins for remote power-management of the managed systems. That allows openQRM to automatically shutdown and power-off unused resources and power them on just when they are needed.

5. Summary

All those abilities of openQRM allows IT-departments to decrease their number of physical systems while keeping, even increasing, the manageability, high-availability and automated monitoring facilities of their services by other, plug-able features for openQRM. This means less systems which are consuming power. It also means that less warmth is produced in the data-center. This again results that less energy is consumed by the cooling-systems of the data-center.

Link to the openQRM website
Link to the openQRM project

1 comment:

Maz said...


Cool Stuff, where get i get some guide , howto setup openqrm on at least to servers. I like to start some lab test soon. I could not find it on openqrm website. Thanks.