Automation and Orchestration: Two Key Steps along the Road to SDI


Ready to know more about that journey along the road again? In an earlier post, I posed the question, “Where are you on the road to software-defined infrastructure?” In that post, I outlined the road to SDI in terms of a maturity model framework with five stages: Standardized, Virtualized, Automated, Orchestrated, and SLA Managed. In this post, I will take a closer look at the Automated and Orchestrated stages.

These stages are very like twin cities on the road to SDI. To get the second stop, Orchestration, you need to get beyond the first, Automation. So we will begin there.

In the Automated stage of the SDI maturity model, IT resources are pooled and provisioned with minimal manual intervention and with the right security in place. In a typical progression, an organization will automate the provisioning of routine tasks like the installation of patches and the deployment of virtual machines because of the variability of managing the existing environments.  Olny thenwill they be able to move on to the automation of resource pools.  With this complete, they can finally automate the application stack with approaches like platform as a service (PaaS) or other application related portions.

There’s one important twist and turn on the road here. To put your organization on the best path to SDI, you need to start with more than a collection of standalone tools that provide automation capabilities like you have already in most environments. You need tools can or already are integrated into an end-to-end orchestration platform.

To achieve this goal of gaining an integrated toolset, you basically have two choices. You can pick out best-of-breed products and do the heavy lifting required to integrate your own custom environment, or you can buy an off-the-shelf solution that is tied to a single vendor’s platform. I explored this topic in an earlier post that asked the question, “Should you take the high road or the low road to SDI?

Regardless of the route you take, automation provides the basic toolset and pooling necessary for orchestration, as well as capabilities for cloud-like provisioning models. So once your automation tools are in place, you’re poised to move on to the orchestration of IT processes.

Orchestration takes automated processes to a new, more intelligent level. At this higher stage of maturity, an orchestration platform optimizes the allocation of data center resources. It collects hardware telemetry data and uses that information to place applications on the best servers, with features that enable the acceleration of the workloads. (WARNING: For the road to be possible, most applications have to be more cloud-aware, a topic I will cover in a future blog).

The orchestration engine takes its direction from policies that you set up at the front end. Following these policies, it places workloads in approved locations for optimal performance and the right levels of security. It also acts as an IT watchdog that spots performance issues and takes remedial actions automatically. Along the way, the orchestration platform accumulates learnings that it will put to use to make even better decisions in the future.

Here’s another twist to keep in mind. We’re in a period of rapid evolution in the orchestration space. This is a young and immature part of the SDI ecosystem. We’re basically at Orchestration 1.0. Things will be quite different as we move to Orchestration 2.0, with more mainstream solutions, and 3.0, the point at which solutions will be firmly established and widely used.

What does this mean to your organization? At a practical level, you need to take a long-term view. Be flexible in your scripting, and create a foundation that gives you choices for the future. Think about your applications and how they can be designed to work in an orchestrated environment. On the road to SDI, people tend to focus on the next stop, rather than the entire journey. Those who focus on the big picture will be in a better position to realize the full promise of SDI.

From a big-picture point of view, automation and orchestration create the foundation for the ultimate destination—the SLA Managed stage in the SDI maturity model. At this stage, SDI allows you to deploy hybrid clouds, control workloads and enforce security through an end-to-end orchestration layer that automatically enforces policies for application workloads and the service levels they require. The SDI environment makes sure the application gets the resources it needs for optimal performance and full compliance with the governing policies.

This isn’t a dream. It’s a very definite designation for enterprise data centers. The key is to find the route that is right for your organization.

Happy travels!

Find me on Twitter @EdLGoldman to share your thoughts.