Is Flexibility Scary, Dangerous, or a Necessity for Infrastructure Automation Success?

One of the things that Infrahub brings to the world of infrastructure automation is radical flexibility in the types and scope of data that it manages, all within a continuous integration lifecycle process that is usable by infrastructure engineers. This stands in contrast to the vendor-defined data models and schemas of many historical products in the field. In fact, it might be tempting for some to cast flexibility as scary, dangerous, inconvenient or messy. But in our view, flexibility is mandatory for long-term infrastructure automation success. That’s why the flexibility built into Infrahub is purposeful. Without that flexibility, you can’t create automation to run your business the way you need.

Why Do Most Infrastructure Automation Products Feature Fixed Schemas?

There are multiple reasons for this phenomenon:

1. The PoV of the original creator. In many cases, the original creators of infrastructure automation tools were approaching things from a particular domain perspective (network, storage, compute, etc.) and that heavily influences the model they develop, because it’s the boundaries of the problem statement they were trying to solve.
2. Vendor-specific tools. Historically, most toolsets were produced by hardware vendors, so it was natural and too easy to build a model that, if it wasn’t exclusive, heavily favored that vendor’s products. 
3. Relative immaturity of infrastructure automation. Historically, with rare exceptions, infrastructure automation teams have lagged far behind dev teams in operating CI cycles for example. This meant that product developers catered to users in many cases that needed and benefited from the automation process being pre-defined and pre-packaged.  
4. Technology barriers– historically, most products were built on SQL databases, where table structures are costly to build, relate, and change from the point of view of index fields, etc.  
5. Product vs Platform approach. When you’re trying to make things easy, and you want to constrain the scope of your own product and software maintenance, and especially when your product vision has a limited scope, then you’re going to build more of a packaged product than a platform. By contrast, a platform has to be highly flexible to accommodate how businesses want to shape their data and match the way they want to deliver infrastructure as services, what infrastructure they have, and all the potential sources of truth that exist in the organization.

Why is a Flexible, Integrative Platform Approach Appropriate Today?

Simply put, the market has matured. There has been a groundswell in automating infrastructure, and many teams are ready to go beyond previous constraints so that they can build to the shape of their business.  

Think of this as analogous to a company that chooses to use Salesforce or SAP. These are powerful, flexible, modifiable platforms that can accommodate a ton of different types of business data. They require investment and work, but they can run complex, evolving business processes, whereas simple tools often cannot.

There are now a lot of companies that have implemented the ground stage of their infrastructure automation. They want and need to move beyond low-level work, and start building a catalog of consumable services from their infrastructure that can drive consistent business performance outcomes:

• Time to market to build a new location, store, PoP, data center, etc.
• Time to update existing infrastructure with new vendors, OS, features, or pieces of infrastructure
• IT and automation staff efficiency
• Consistency and predictability of configuration changes for uptime, continuity, security and compliance purposes

These businesses must often accommodate a lot of different “sources of truth” for infrastructure data, including home-grown databases, vendor-specific configuration management databases, CMDBs for physical inventory and asset tracking, network SoT, security policy managers, etc. So they need flexibility to integrate all of that. 

And they need a platform that can allow them to capture, maintain, and evolve their service designs, bring together all their infrastructure data, translate between the two, and keep rendered data persistent—all within a CI process that allows everything to be versioned and lifecycled from design through to pushed configs. 

When you have these needs, flexibility isn’t scary or dangerous. It’s exactly what the doctor ordered. These organizations understand that a platform approach requires some building, but that the right platform makes what they build sustainable, reusable, and performant.

What’s Our Prediction for Automation that Tries to Grow Without Flexibility? Pain.

In the Enterprise Management Associates (EMA) Network Automation Report from March, 2024, there were some eye-catching statistics. One was that 90% of the survey participants reported that they had a network source of truth (NSoT) product deployed, but only 20% considered it to be successful. Another stat was that overall, only 18% of participants considered their network automation to be completely successful.

When you don’t have flexibility, you leave a lot of things out of your process, or you try to bolt them together with limited staff time to develop and maintain. Then over time, complexity kills the cat. It’s a tale as old as time, or at least as old as infrastructure automation.

Infrahub: Purposeful, Powerful, Platform Automation That Works for Any Infrastructure

There are good, mature solutions for mature application development automation and even for spinning up cloud automation for dev, test, or production software deploys. *Cough Git *Cough.

But there has been a glaring lack of a platform that gives the flexibility and power needed to build a sustainable, full infrastructure automation lifecycle, from data through deploys. This is what Infrahub delivers. It is purposefully flexible to meet the needs of teams that are ready to mature and scale their infrastructure automation. Infrahub unifies all your data (without needing to be canonical), pulls all your data into a CI process, and deploys to all your things using your tool of choice.

There is a very good place for more constrained, pre-defined automation tools. If you’ve never done any automation, are just trying to level up from scripts, or have limited needs, that could be a good place to start. And thankfully there are open source, vendor-neutral tools that can help you with this part of your journey.

But if you’re ready to level up, ready to embrace flexibility and enable your automation to drive your business goals, then help yourself and check out Infrahub on Github, or hit us up and we’ll walk you through how it works live.