The provisioning flow

The provisioning flow consists of:
• Detecting all the resources in the POD
− Catalog compute, storage and network resources
− Rack Scale Architecture Controller decides what constitutes a storage node. The Controller could also assign one or more compute nodes for front end storage processing in storage stack
• Selecting nodes for primordial storage
• Installing Primordial storage stack
• Assigning the remaining Storage Nodes available for assignment to SDS controller (e.g., Cinder and Manila in case of OpenStack, or CoprHD open source SDS controller)
• Assigning switches that are available for assignment to SDN Controller (e.g., Neutron)
• Assigning Compute Nodes that are available for assignment to SDC Controller (e.g., Nova)
• Composing Compute Nodes with boot volumes coming from Primordial storage stack (OS and local data store in the boot volume per Compute node)
• Activating OpenStack Heat template to define server host resources:
− CPU type
− Performance SLA
− Memory, storage, and network
− Advanced – workload template (for Intel® Rack Scale Architecture 2.0)
• Composing a server and adding to an Aggregate namespace
• Creating a secure environment for booting. Options are:
− A separate VLAN for copying image to storage volume mapped to host
− A separate secure pool
− A boot namespace
• Mapping the Glance image to composed host
• Mapping the storage volume for boot
• Copying the image securely to volume and boot
• Allowing Fuel to provision OS, IP address and OSS agents and add to Nova database OR
• OR TripleO