BonFIRE logo and link to main BonFIRE site

Previous topic

Restfully

Next topic

Contextualisation

This Page

Overview of ComputeΒΆ

BonFIRE offers compute resources to be able to deploy VM Images on the different testbeds. A compute resource MUST be deployed with a network resource, which could be one of the default networks provided in BonFIRE or one that you create yourself, as discussed in Overview of Networking in BonFIRE. A compute resource can, optionally, also be deployed with an additional storage resource if the storage space provided with the VM image is not great enough (maximum 10GB currently). For more information about this, see Overview of Storage.

When compute resources are deployed, there’s some Contextualisation that occurs to configure details such as the hostname, public IP (if any) and bonfire credentials. Such things are provided automatically by BonFIRE, but experimenters can also provide parameters such as aggregator IP (if any), metrics (if any) and post-install scripts. The contextualisation element is generic, so that any key-value pairs could be defined. The contextualisation variables are written to /etc/default/bonfire, so you can source this file to access the variables within your experiment VMs.

To support large-scale experiments, BonFIRE offers On Request Resources, which can be reserved in advanced and used as the ‘basic’ compute resources. For more details and how to deploy the compute resources, see Deploying Compute Resources in BonFIRE. It is now also possible to deploy resources on Amazon EC2 within a BonFIRE experiment.

As a facility to conduct empirical experiments, BonFIRE strives to provide maximum control of how you define and where you deploy the compute resources. To deploy the compute resources, BonFIRE offers two ways of specifying the Instance Types , following either a ‘small’, ‘medium’, ‘large’ type of taxonomy or your own custom template based on CPU, VCPU and RAM. You can also specify the physical host or cluster that the resource should be deployed on.