Compute categories and tags

Categories and tags are designations applied to compute nodes that allow you to apply constraints or filters on your compute resources.

A compute category is a descriptor of a node, and each compute node in the system must belong to exactly one compute category. Compute categories are used to organize compute resources for allocation to virtual datacenters. Compute categories are useful for organizing a large infrastructure to define specific stack settings like user management or other application settings, such as computing power or memory requirements. If needed, you may also restrict an application instance to run only on nodes in a specific compute category.

Compute tags are additional metadata that can be applied to nodes. While each node must belong to exactly one category, a node may have any number of tags. Because a compute node can have many tags, this helps you to be more specific in your designations. Compute tags are used by application instances and nodes to determine which nodes are eligible to run an application. If an instance lists compute tags in its resource profile, only nodes with all of those compute tags will be able to run that instance.

In the ThinkAgile CP environment, when an application runs, it requests a compute resource with certain tag and category designations for the computing resources needed instead of requesting a specific hardware server on which to run. Traditionally, applications are tied to specific hardware; however, in the ThinkAgile CP platform, an application is not required to be assigned to a particular compute node, which allows for flexibility in application deployment and ensures a hardware-agnostic approach. Applications do not have to specify hardware to run; they can specify computing power or memory requirements based on categories and tags.

For added stability, another compute resource can be chosen automatically with the same category and/or tag designations to serve the application in the event the compute node goes down.