Infrastructure-as-Code (IaC) tools have been used by IT professionals for over a decade to help define and deploy cloud infrastructure.
The promise of IaC is that infrastructure can be declaratively defined in code, enabling an organization to deploy the same infrastructure again and again. In the modern cloud world, however, the complexity of different cloud providers’ services has made it difficult to define infrastructure in a way that’s portable across providers. It’s a challenge that new startup FluidCloud is aiming to solve. The company is launching out of stealth today with $8.1 million in seed funding led by Unusual Ventures.
“We’re actually translating the infrastructure as code from one dialect on one cloud, and being able to port it and transform it into a dialect of another cloud,” James Bayer, co-founder and CPO of FluidCloud, explained to Network World.
Where existing IaC tools fell short
The startup’s genesis came from real-world frustrations.
Bayer had previously been at IaC pioneer Hashicorp, which was acquired by IBM for $6.4 billion earlier this year. Bayer recounted that after Hashicorp completed an acquisition in 2023, it took time and money before they were able to port the technology from Google to AWS.
“We spent tens of millions of dollars on an acquisition, and we couldn’t sell it for six months because we had to wait for porting to happen,” Bayer said.
Harshit Omar, co-founder and CTO of FluidCloud, said that while IaC is widely adopted, every cloud provider has its own rules about how things can be deployed. “The way you work and deploy things in AWS, the way you deploy things in Azure, that is totally different,” Omar said.
Patented algorithm solves factorial scaling challenge
Cross-cloud infrastructure mapping traditionally creates what Bayer referred to as factorial complexity.
Each cloud provider requires unique translation paths to every other provider. Six clouds demand 30 distinct mappings. Adding a seventh cloud requires six additional translation paths.
FluidCloud developed a patented approach that converts factorial scaling into linear complexity. The system creates universal intermediate representations that translate to any supported cloud dialect. The platform currently supports AWS, Azure, GCP, VMware, Vultr and Nutanix with additional providers under development.
FluidCloud’s discovery system calls cloud APIs directly to inventory existing infrastructure across all supported platforms. The engine can scan 50,000 resources in 15 seconds compared to hours required by competing tools that perform similar functions, FluidCloud asserts.
Real-time translation preserves network architecture
FluidCloud’s core engine translates infrastructure between cloud-specific resource dialects while maintaining network topology and security relationships.
In a demo, Bayer showed how the FluidCloud system was able to map an AWS to Vultr translation, converting EC2 instances, VPCs and security groups. The translation system handles complex resource mappings automatically. AWS t3.small instances become equivalent Vultr compute configurations based on CPU and memory specifications. Security groups translate to firewall rules with identical port protocols and source restrictions.
Network isolation transfers intact across cloud boundaries. Private address ranges and security group rules permitting port 5000 access replicated precisely from AWS to Vultr during the demo as well.
It’s important to note that FluidCloud is not actually building its own IaC technology; rather it is using existing technology, specifically Hashicorp Terraform.
“We are using Terraform as the lingua franca of DevOps. We are not trying to introduce our own proprietary language,” Bayer said. “What we do have is proprietary capabilities of translations for TerraForm.”
AI integration extends platform capabilities
Future development includes AI-powered automation agents for failover scenarios and outage response.
The team also will be working on a multi-cloud MCP (Model Context Protocol) server providing natural language infrastructure management across providers. The multi-cloud MCP server will differentiate from single-provider solutions by offering cross-cloud capabilities through natural language interfaces.
The big source of traction for the company, at least initially, is expected to come from existing VMware users.
“Our core focus is the desperate buyers, which are primarily the VMware customers, which are planning to come out of VMware and trying to move to other cloud providers,” Sharad Kumar, co-founder and CEO of FluidCloud, said.
Source:: Network World