Breaking Free from Virt Mangement

One of the next steps for ManageIQ is to begin understanding how to manage non-virt providers types. Virt Providers in this context really refers to Private Infrastructure, Public Cloud, and Private Cloud providers.

Two examples of non-virt provider types would be:

  • Containers
  • Storage

Of course, there may be virt elements to these other provider types. But, the goal is to understand these providers free from the context of virtualization. ManageIQ needs to understand what’s interesting to the user about these provider types and how to surface that information most effectively to the user.


An example of a Container provider is Docker. It could also be that Openshift is a container provider. Or, it may be that we treat Openshift as an example of PaaS, and let it be altogether different. The problem at this point is in understanding what’s interesting to report about either concept.


Ceph, Gluster, NetApp…just a few examples. Do we also want to start considering Amazon S3 as a separate storage provider? What about Dropbox?

If we only consider storage as how it’s used in context of virtualization (image management, external storage for VMs), then this constrains the scope considerably. But, I’m not altogether certain that’s a good thing.

I would also add network to complete the IaaS triptych : Compute, Network, Storage.

I agree with Fabien. Has thought been given to the best way to integrate network orchestration (code directly to underlying fabric (OpenFlow, Open vSwitch), SDN controller North-bound APIs (OpenContrail, OpenDaylight, NSX, Programmable Flow, ACI) or integrate with network management tools (CloudBand, APIC, Cyan, Tail-f)? Or a mix of all?


@fdupont, Good addition!

@dneary, I think the approach would likely be to integrate with the management tools. But, honestly, probably a mix of both. I would see the integration with orchestration maybe happening on-site instead of directly in the product. Possibly? Just a hunch right now, though.