Adoption

Start from what you already run.

There is no blank page and no migration project. Fractal Cloud reads the infrastructure you have today, turns it into something you can reuse, and brings it under management as it is.

Two Ways In

You start from the assets you already have.

There are two of them, and most organizations use both. Whichever way a system arrives, it lands in your library and is governed from then on.

Approach one

Scan and Import

The platform reads a live account, shows you what is there, and brings it under management as it is. Four steps.

See how it works →
Approach two

Your Terraform

Read and rebuilt as a Fractal, or kept as it is inside a component the platform governs. You choose per module.

See how it works →
Approach One · Scan and Import

Four steps, and your estate is in.

Connect a cloud account, scan it, read the report, group what was found into Fractals. At the end everything you already run is visible in Fractal Cloud, exactly as it runs.

Step One

Point it at your infrastructure.

You add an agent for the environment you want to read: an AWS account, Azure, Google Cloud, OCI, a sovereign provider or your own datacenter. You give it credentials, choose the scope, and it is ready.

  • What runs in your own datacenter is read the same way as a cloud account
  • The scan changes nothing while it reads
  • Several providers can be connected side by side
Adding a cloud agent for VMware, credentials verified and ready
Step Two

Read the whole stack.

The scan walks compute, storage, network, big data, edge and identity in turn. Every resource found is written to a live log, so you can follow it as it happens.

  • It reads only, and changes nothing
  • Minutes, not a project
  • Several accounts can be scanned side by side
A scan in progress across compute, storage, network, big data and identity
Step Three

Read the report.

A discovery report with discovered, recognized and unrecognized counts

Three numbers: what was discovered, what the platform already recognizes, and what it does not. The last group does not block anything, and the table lists every resource in detail.

On its own this is already useful. Most organizations do not have one accurate picture of what runs across their datacenter and their cloud accounts, and this produces one in minutes.

Step Four

Group what was found into Fractals.

Discovered resources grouped into Fractals, with two still unassigned

The platform proposes a grouping, and you correct it. Anything it could not place is dragged into the Fractal it belongs to, and you can create new ones. This is where your knowledge of the estate is captured.

  • What comes out are named Fractals, not a flat inventory
  • Nothing is deployed or changed at this point
  • You can change the grouping at any point
Once It Is In

A design you can reuse, or the estate as it runs.

At the end of the four steps you decide what each system becomes. The difference matters.

Import as a Fractal

A reusable design

The architecture becomes a versioned design you can review on a canvas, edit if you want, and deploy again as many times as you need.

Import as a Live System

The running estate, governed

What is already running is brought under management as it is, so it gains versioning, reconciliation and controlled updates without being recreated.

Both routes end in the same place: your own architecture, versioned, in your library. Deploying it somewhere else is a separate decision, and a separate story, see running it where it has to run.

Approach Two · Your Terraform

The Terraform you have keeps running.

Scanning a cloud account is one way in. The other is the code you already wrote. Two routes, and neither throws anything away.

The Terraform you have
network.tf cluster.tf iam.tf + 37 modules
Route one · read and rebuild
It becomes a Fractal.

The code is read and turned into a blueprint with an interface, so it can be reused, versioned and deployed on any target.

blueprint interface v1.0
Route two · keep and wrap
It stays Terraform, inside a component.

The module runs as it is, wrapped in a component the platform governs. Teams that write HCL keep writing HCL.

component wrapping module.tf
Your library
governed the same way
Whichever route a module takes, it is versioned, policy checked at deploy and continuously reconciled from then on.
Where This Helps

Four situations this was built for.

Leaving a datacenter

You know what runs on your own hardware, roughly. Reading it properly and rebuilding it on a cloud is usually the longest part of the move. Here it is the same design, deployed again.

Inheriting an estate

After an acquisition, or when a team changes hands, you get accounts nobody can fully describe. The scan gives you the picture, and the import gives you control of it.

Proving where things run

An auditor asks what exists and where. An inventory produced by a scan, kept current afterwards, is a better answer than a spreadsheet maintained by hand.

Standardizing what exists

The architecture your best team already built becomes the design every other project starts from, instead of a repository the others copy and slowly change.

Begin with one environment

Connect one provider, run one scan, and see your own infrastructure come back as something you can deploy again. Everything after that is your decision.

Book a demo