“Managing network infrastructure around the globe is a pain,” said Damien Garros, OpsMill CEO and co-founder, in an interview with TNS. “I had to build my own software. There was no other option.”
Garros’ new startup has released this work, in a new Infrastructure as Code (IaC) platform, currently in beta, that promises to automate infrastructure deployments through a GitOps-based Infrastructure as Code approach.
Why couldn’t system admins enjoy the same level of automation enjoyed by the developers?
Infrahub was born from frustrations Garros kept encountering with infrastructure and network management while building IT automation systems for a variety of organizations. All the configuration info, for managing a virtual private network, say, or for setting up a load balancer, is usually not captured in any sort of systematic way, beyond engineer notes in a RTFM text file somewhere.
The tools that are available for managing such configuration data are clunky, cumbersome and costly. And they have serious technical limitations, most notably due to their inflexible data models.
“To automate, you need good data,” said Raphael Maunier, OpsMill chief operating officer and co-founder, in an interview with TNS.
In terms of IaC, the source of truth when it comes to complex systems management is always data. So, why not put it in a database?
How InfraHub Stores Configuration Data
A more supple data model is key to better automating IaC, the OpsMills execs insist.
To date, infrastructure management practices have largely ignored the GitOps-approach, where changes made in code (captured in git) are automatically deployed to production, with all changes automatically captured in version control.
What’s missing with GitOps, in terms of its potential use for infra management, is a structured data model, essential for capturing the many inter-linkages across components.
Using git itself to hold the configuration data has its limits: It doesn’t have a good query engine, it doesn’t offer a schema, and it can be difficult to use in this capacity, Maunier explained.
Try managing your IP addresses in git. It’ll be a mess, Maunier said.
To aid in data management, Infrahub uses a graph database, from Neo4J, which is ideal for capturing inter-relationships across many components, and can also easily support an extensible, ever-evolving data model.
Between git and the graph database, Infrahub captures all the configuration data needed for even a complex system.
This “intent” data describes how a system should be assembled, and Infrahub can feed this data to HashiCorp Terraform or Red Hat Ansible or other IaC deploy menttools.
What are the Benefits of Infrahub?
The Infrahub platform offers versioning. Multiple admins can create a branch of the infrastructure database, add a new component, then merge the update back into the database, automatically generating configuration files in the process, accessible by API.
The software can also do continuous integration validations to ensure operations continue with the new components installed.
Another benefit is compliance. Multinational companies have to abide to various laws around the globe. With Infrahub, components can be tagged by which countries they are in, and their configuration information can easily be collated for compliance inspection.
The software is open source, and can be deployed on premises. And the company is working on a managed cloud version as well.
A venture-backed startup, OpsMill is backed by Serena Capital, Partech, OVNI Capital, Kima Ventures and Better Angle.
The post OpsMill Infrahub Automates Infrastructure as Code with GitOps appeared first on The New Stack.
The missing ingredient to automating infrastructure deployments is a better, more extensible data model, the OpsMill founders explain.