A Pragmatic Approach to Automation
We are nearing the top of the hype cycle for data center automation. There is much discussion regarding how necessary it is and how it will be the future, and there are many different tools and products available, all implying a sense of urgency to get on the bandwagon. Despite the noise level, there is one area that is often neglected when discussing automation–HOW to automate and the fact that it's really a journey and not an actual destination. The best method for adoption is to automate at your pace based on your organization's unique requirements. Every organization has different needs, skillsets, environments, and resources. Rebuilding a plane flying across the globe at 36,000 feet while you’re on it isn’t the most appealing option for anyone, yet, that is what many vendors are suggesting when they leave out the gory details of “how” and talk about all the bright, shiny objects of what’s possible.
Digital transformation is forcing our hand
We are in the midst of retooling our data centers in order to support digital transformation, which marks a radical shift in the way we do business. The data center is the digital central nervous system of the business and must be modernized so it can rapidly adapt to minor course corrections or major course changes at a speed that is unheard of in the recent past.
While modernizing the data center infrastructure, we also need to lay the groundwork for automation. But let's be realistic about this–we aren't going to automate everything day one and walk away simply. In keeping with our transportation theme, we can see an analogy between data centers and aircrafts (just stick with me here for another minute or two). When aircrafts first came into existence, they were flown 100 percent manually. Over time, levels of automation were added to provide autopilot functionality, flight management controllers, crash avoidance, and numerous other features that have evolved the pilot to a person that is the expert in the cockpit, monitoring the flight but with little to no intervention unless necessary. The evolution of the pilot also included an evolution in how the pilot was trained. They still need to know how to fly the aircraft, but in addition to that, they must also understand all the components of the aircraft’s automation capabilities. Starting to see where we are going here?
The path to full automation
Consider that automation is not merely for deployment but should be thought of in the full life cycle of the data center: deployment, validation, troubleshooting, and remediation. With that, we each need to find our path to automation and see where we can jump in. The stages of automation are as follows:
Stage 0 – No Automation: requires human intervention for every configuration change in the data center across all components of networking, servers, storage, applications, and services
Stage 1 – Automation assist: allows simple tasks to be automated but does not necessarily require customization by the end user. This form of automation is targeted at a single domain within the data center such as deploying a virtual server but does not automate its networking or storage services
Stage 2 – Customized automation: builds on Stage 1 with the ability to customize the workflows to suit the needs of the environment better. This is still within a single domain but adds workflows that perform multiple actions such as creating a VLAN, default gateway, and port channel for a specific server’s network connection
Stage 3 – Cross-domain automation: integrates the different functions across the data center in automation workflows that cross some or all networking, servers, storage, and services. This level of automation can build on Stage 1 and Stage 2 but is not dependent upon either of the previous stages.
Stage 4 – Full automation: removes the need for human intervention with the use of Machine Learning and Artificial Intelligence allowing the data center to adapt as needed based on the analysis of data and information collected in context.
Each of us will likely start and stop at different stages along the journey. That's the beauty of automating at your pace and having a choice. This is all about becoming more effficient in all the right areas to make your data center an asset and not an anchor. Beginning the journey is most important, and that starts with upping your skillsets in terms of automation tools that use programming languages such as Python and Go which have tremendous mindshare. Be wary of tools that lock you into a single vendor as that limits your abilities to adapt in this wonderous multi-vendor world that is only getting bigger and growing faster. And remember, "don't work late, automate."
Recommended For You
Most successful IBN deployments focus on the network verification process. Not only is it safe, it also can easily integrate into existing networks and workflows.
Continuous monitoring and baselining of net performance monitoring metrics can reveal problems before users do and prevent complaints on performance degradation.
It's time to move past some common misconceptions and fears about SD-WAN. Here are three common myths you can ignore.
As the routing protocol that runs the Internet, BGP is a key piece of the puzzle that helps you understand how your customers get to you.
From a network planning and design perspective, manually created diagrams drawn by a human architect will continue to be the go-to method for years to come.