There’s no such thing as a network automation fairy.
That’s not to say that people haven’t tried to build one. Terraform, Ansible, and similar platforms dramatically simplify network automation by providing a framework to build with. Anything that makes network automation simpler is cause for rejoicing.
Yet even with the leg up these platforms provide, there’s still a lot of grunt work involved in network automation. You have to tell the automation platform what you want it to do, and where you want it to apply the magic. In the end, this is something that only an in-house network team can do properly. Only network operators have the knowledge of assets and processes to make automation a tangible reality.
Scoping out network automation
At least half of the grunt work is figuring out where automation will be applied – which devices or virtual assets you need to touch as part of a process. Terraform, Ansible, and other automation platforms offer no help on this front. They will orchestrate a process once you tell them where to do it, but they only know about the assets you point them to.
This is where NetBox adds a ton of value. With NetBox, you’ve got a pre-populated inventory of physical and virtual assets that integrates with your automation platform of choice. All you have to do is orchestrate the connection between assets and processes.
But what happens when you don’t have a network source of truth to draw from? Is it possible to just populate information about your network on the fly? What’s the level of effort involved in doing it yourself?
Network automation without a network source of truth
Let’s say you want to perform a single, very simple task using network automation, and you don’t have a network source of truth in place.
Define the scope, collect data
First, you’ll need to figure out which assets you need to touch with the automated process. That means collecting information on everything in your network that’s in scope (physical devices, IP addresses, VMs, etc.), and ruling out everything that isn’t.
Then you’ll want to figure out how your automated process will interact with those assets. How do they connect to the network? Do you have the credentials you need to make an automated change? What do the device configurations look like? How will you actually make something do what you want it to do?
Create a data store
Once you’ve got an inventory of assets and a sense of how they interact with the network, you can start building out the automation process in a platform like Terraform or Ansible.
This is where it gets ugly. As part of the workflow, you have to draw on your list of in-scope devices and tell your automation platform where to apply its logic. That means pointing the automation platform to some kind of organized data store which has all the information ready to go.
Technically, you could compile all that data in a spreadsheet, TXT, or CSV file. It sounds easy enough, but when you think about confirming all of that data, keeping it updated over time, and accounting for scale, it’s actually a huge lift.
There’s an added complication of telling your automation platform how all the assets in your data file are related. Spreadsheets, TXT, and CSV files simply aren’t set up to do that. Representing the linkages between assets requires coding.
Integrate with an automation platform
That brings us to yet another challenge with using a spreadsheet, TXT, or CSV file to hold all your inventory data: the ingestion step. Automation platforms need some kind of integration to know how the inventory file should be processed. They aren’t set up to automatically recognize spreadsheets, TXT, or CSV files as inventories. This requires both coding (to create an integration between the inventory file and the automation platform) and configuration (to guide processing of the inventory file once its contents are in the automation platform).
Keep in mind that all of this work would need to happen for every automation task on the network. For every process you’d need to compile a unique spreadsheet, TXT, or CSV file, ensure its accuracy over time, and integrate it somehow with the automation platform you’re using.
That ends up being a lot of effort. Even more, it’s a lot of specialized effort. To integrate a non-standardized source (or sources!) of truth with an automation engine, you’d have to become a coding expert, delving deep in the weeds of architecture elements like parsing engines. In other words, you’re building (and maintaining) custom plug-ins yourself. Multiply that effort across all the processes and systems you want to automate, and it quickly becomes an insurmountable task. The complexity is just too much to handle.
The easy button: a network source of truth
Is it possible to deploy automation without a network source of truth? Yes, it is technically possible. For every automated process you can do the work of defining the scope, gathering asset data into a single data store, coding the integration of that data store with an automation platform, and managing configurations within the automation platform.
Yet doing all of those things for every automated process – and keeping it all up to date with changes in the network and the automation platforms themselves – is an exhausting prospect. Scale is half of the logic behind network automation, yet without a network source of truth it’s almost impossible to achieve.
This is why network automation drives are floundering. Gartner estimates that as of early 2022, only 8% of enterprises automate network activities. Despite all the hype (and tools!) around network automation, not many companies are actually doing it on the ground.
NetBox Cloud gives network teams at least one easy button on the long path to network automation. NetBox Cloud compiles relevant information about your network and dishes it out to automation platforms in a format they can ingest and use. It eliminates the need for coding and integration work – all you have to do is manage configurations on your automation platform.
NetBox Cloud also gives network teams the ability to scale any automation drive by re-using existing asset information across a variety of use cases. That also means keeping a single data store up to date rather than constantly updating process-specific inventories. Since it handles every element of your network – physical or virtual, from the largest data center to the smallest branch office – NetBox truly offers an inclusive inventory of everything you’d want to automate.
Simply put, a pre-built, comprehensive, fully integrated network source of truth is a necessity for any network automation drive. NetBox is used by so many enterprises of all shapes and sizes because it eliminates the complexity and effort associated with hacked-together, difficult to maintain data sources. It’s not just “a” network source of truth. It’s “the” network source of truth.
Looking for more information about getting started with network automation? Read this tutorial. To check out the open NetBox demo, click here.