Infrastructure-as-code company Platform Engineering Labs has announced the launch of formae, an infrastructure-as-code (IaC) tool for engineering teams managing cloud-based infrastructure. Aimed at DevOps and platform engineering teams, formae is designed to provide a unified, code-centric approach that helps platform engineers focus.
“You can do infrastructure changes based on code, in any granularity, without having to see any detail if you just changed a single property on a single resource,” said Pavlo Baron, co-founder and CEO of Platform Engineering Labs. “This is important because focus is what lets you succeed.”
Baron believes the tools available for platform engineers are not up to the task. “Everybody tells us our shell scripts are bad, our Bash is bad, our tools are bad. Well, guess what? I mean, Terraform is no different. You need to dance around to have a proper loop in Terraform,” he says.
Rather than using the domain-specific language used by Terraform and OpenTofu, HCL, Formae uses the PKL configuration-as-code language to interface with the system. Originally developed by Apple, the language has been open-sourced with the Apache 2.0 license. Platform Engineering Labs considered using the open source KCL language, hosted by the Cloud Native Computing Foundation, but found it to be “less advanced” than PKL.
Formae uses PKL in two modes of operation: reconcile mode, and patch mode. In reconcile mode, customers define their desired environment with PKL code (either created by formae by extracting the running infrastructure configuration) and formae reconciles this “source of truth” against what is currently running in production, reconciling the two. In patch mode, customers focus only on what needs to change, submitting PKL code to formae which then checks that the changes are sensible before patching the running environment with just the minimal changes provided. Both modes can be used simultaneously, depending on a customer’s preferred workflow for the situation.
Formae currently targets infrastructure in AWS, but plans to support other cloud providers based on customer demand. “We can look into stuff like OVHcloud in Europe, because Europe has this massive push towards sovereignty, and we can actually piggyback this movement,” says Baron.
“We want to democratize access to infrastructure management,” says Baron. “That’s our goal, and we want to make it work for everybody in the way they actually expect it to work.”
How Open Is Open?
Platform Engineering Labs has chosen the new and somewhat niche Functional Source License (FSL) for formae. The FSL is “source available” license that makes source code available but restricts how the software can be used as an attempt to guard against free riding. The FSL is an adaptation of the Business Source License (BUSL), created by MariaDB and used by HashiCorp to relicense its products prior to being acquired by IBM.
Licensing is a touchy subject in open source circles. Several companies that had made their software available under open source licenses came to believe that free riding was undermining their ability to maintain the software and chose to change the license to one with restrictions on activities the sponsoring companies saw as unfairly competitive. The relicensing of Terraform led directly to the creation of the OpenTofu project, a fork of Terraform that retains an open source license. MongoDB’s attempt to have the Server Side Public License (SSPL) accepted as an open source license was rejected by the Open Source Initiative, the generally accepted arbiter of what counts as open source, and what does not.
It remains to be seen how accepting customers will be of FSL licensed software. Adoption of the BSL remains rare, and some companies, such as Elastic, have chosen to reverse their relicensing decision and return to open source. Source have told me they have concerns about the way security patches are handled under licenses like the BSL and FSL and the interaction with the conversion date to an open source license. Yet the sustainability of open source projects remains a significant issue, which I have written and spoken about extensively.
Exciting Times in Platform Engineering
Platform engineering is enjoying a surge in attention as developers increasingly realise that managing platforms is a separate discipline with its own unique challenges. System administrators, meanwhile, are embracing the tools and experiences that developers have long enjoyed. The preferential treatment afforded to developer experience has meant infrastructure systems and tooling have been somewhat neglected. Platform Engineering Labs joins other companies such as Pulumi and System Initiative that are hoping to improve the experience of platform engineers as much as developers.
The recent AWS outage has also reminded people that efficiency can come at the cost of reliability and increased risk. Yet hedging against downtime at any one vendor, be it a network provider, cloud provider, or other single point of success can be complex and costly while introducing risks of its own.
Quality tools for managing infrastructure across providers are an important part of keeping the cost and complexity manageable while also accepting a decrease in efficiency to guard against risk. Major environmental changes, such as Broadcom’s acquisition of VMware, IBM’s acquisition of HashiCorp, and new geopolitical tensions are inviting customers to rethink how they manage the infrastructure that supports the critical systems that underpin our modern, Internet-connected society.
Disclosure: Pulumi is a PivotNine client.
