Most people working in IT today have probably heard loud and clear that hybrid multicloud is the new “new normal.” In this blog post, I’ll share an example of how you can use IBM Power® Systems technology in a hybrid multicloud environment with management software available in Red Hat® OpenShift® Container Platform and IBM Cloud Pak® for Multicloud Management.
By now we’ve all either read about, experimented with or implemented workloads in a container environment. These new container technologies can work together with virtual machines on Power Systems within your data center and on multiple public clouds and still be managed through a single control system. So let’s jump in and take a quick look at how you can get real business value with software that lets you integrate these technologies.
Let’s talk speed and agility
You’ve just completed that design workshop for modernizing your business application. The new design may have components that can take advantage of the new container technology but also need to integrate with existing VM-based applications. IBM Cloud Pak for Multicloud Management running on Red Hat OpenShift can provide a single view and control into a truly hybrid multicloud environment.
Let’s focus on a specific feature that can help speed the development of our new hybrid application. Within the Cloud Pak for Multicloud Management exists the Terraform and Service Automation Module. This module provides the following key capabilities that can assist you in development of a hybrid multicloud solution:
- Can deploy VMs to private or public clouds
- Can deploy VMs to different cloud architectures, one of which is OpenStack — by providing an OpenStack interface, you can easily connect to your in-house PowerVC environment or to a public cloud that hosts Power architecture, like IBM Cloud. The product has built-in connection templates for over a dozen of the most popular types of clouds
- Can orchestrate a workflow that can combine the capabilities described above to create deployments into multiple clouds in a single action
- Can integrate with the Multicloud Management catalog, which can then facilitate both container and VM-based deployments from a single view and control point
Now, let’s look at a specific deployment example…
5 steps to a new hybrid app
I have a new application that will ultimately run in a Red Hat Enterprise Linux® (RHEL) OpenShift container in a public cloud, which will request data from a Power VM (AIX, IBM i or Linux) running in my private data center.
Step 1 – Let’s create a connection profile to use our on-premises private cloud (that is, PowerVC) as a target for deployments from the Cloud Pak Multicloud Management Terraform and Service Automation Module. It’s pretty straightforward and requires only a few connection and authentication parameters:
Step 2 – Now we need to create a Terraform template that describes the image on PowerVC where the newly developed VM is sourced. Included in the Cloud Pak is a graphical tool called “Template Designer,” which has many tailored templates that can be used as a starting point. In this template, we also define deployment attributes that can either be hard coded or left open to be specified at deploy time — items such as CPU, memory, network and ssh-keys. The designer tool is integrated such that once finished, you can push the template into the Terraform and Service Automation Module. You can find details on using the Template Designer in the IBM Knowledge Center.
Step 3 – Once the template reflecting the VM image and deployment attributes has been created, it can be used to create a “service.” This is where you have the option to orchestrate other activities as part of the deployment. An example might be integrating with a DNS Network IP Registration product to obtain an IP address and hostname, or generating an approval process, or an email notification. Maybe you need 2 VMs deployed together for high availability. There are multiple prebuilt templates available in the tool to allow orchestration steps in a graphical workflow. You drag and drop the action in sequence, and then fill in the specific parameters for that action. The workflow could also include a decision tree that changes the flow based on action outcomes.
Step 4 – Now that we have a service defined, it can be published to the Multicloud Manager Catalog. This is done by simply clicking on a publish button and providing a little information on how you’d like the service to appear in the catalog.
Step 5 – Having this new VM-based application exposed into the Multicloud Manager catalog now makes it visible right along with our partner container and many other services. It’s now possible for various personnel to deploy both the VM to your private PowerVC-based cloud and your container workload into a public cloud.
Now that you have the initial setup, you have the flexibility to allow multiple teams to deploy these workloads over and over, and in ways that make the most sense. Maybe you don’t have enough computer or storage resources in your data center and need to quickly go into the public cloud. You can import your VM image into IBM Cloud PowerVS, add the connection information and modify your target within the service deployment. Maybe you’d like to develop that container-based workload on premises. You can use Red Hat OpenShift running on Power Servers within your data center to keep everything internal until you’re ready to move it out to the public cloud. The flexibility is within your control and with the same tools.
Republished from IBM IT Infrastructure
By Jeff Benson
September 1, 2020