Deployment Models

Solid's three deployment options — SaaS, Private Cloud, and Hybrid — and how to choose based on your security posture and data residency needs.

Solid offers three deployment options. The right choice depends on your security posture, regulatory requirements, and data residency needs.

Can't open inbound connections to your warehouse? If your firewall policy prohibits any external inbound connections — even from an allowlisted IP — see Data System Collector. The collector runs inside your network and makes only outbound connections, and is compatible with all three deployment models below.

SaaS (Solid-Managed Azure)

What it is: Solid's platform runs fully on Solid's Azure subscription. You connect your data warehouse to Solid via the DB Agent (outbound-only connection from Solid to your warehouse, or configured by you).

Includes:

  • All compute, storage, AI services, and databases managed by Solid
  • Dedicated SSO integration layer and operational logging/tracing (external dependencies)
  • Azure AI Foundry LLM services (always Azure)

Your responsibilities:

  • Provision a DB Agent in your environment (or allow an outbound connection from Solid)
  • Configure an IP allowlist on your warehouse for Solid's static Azure IPs — see Solid's Static IP Addresses
  • Provide service account credentials for the DB Agent

Best for: organizations without strict data residency requirements; the fastest path to production.

Private Cloud (Customer-Hosted)

What it is: The entire Solid platform is deployed inside your own cloud environment — Azure or AWS.

Includes:

  • Full platform deployment via Terraform IaC
  • Supported cloud providers: Azure, AWS — Google Cloud Platform (GCP) support is in development
  • LLM: always uses Azure AI Foundry services (Azure AI Foundry connectivity must be provisioned even on non-Azure deployments)
  • Dedicated SSO integration layer and operational logging/tracing (external dependencies)

Your responsibilities:

  • Cloud environment provisioning (Azure or AWS today; GCP in development)
  • Terraform execution and infrastructure management (with Solid's support)
  • Azure AI Foundry provisioning (even if your primary cloud is not Azure)
  • Network security group and VPC/VNet configuration

Best for: regulated industries (banking, insurance, healthcare) with data residency requirements, strict security policies, or internal policy against SaaS for AI workloads.

Hybrid — Control and Data Plane

What it is: Solid's platform is split into two planes:

  • Control Plane — runs in Solid's Azure SaaS subscription (management, orchestration, model engineering UI)
  • Data Plane — runs in your cloud environment (all data-at-rest, all data processing)

Key properties:

  • All data at rest is stored in your data plane
  • Data can be shared with the control plane: encrypted, just-in-time (JIT), and volatile — not persisted in the control plane
  • Solid manages operational complexity while data sovereignty stays in your environment

Best for: organizations that require data residency guarantees but prefer to delegate platform operational burden to Solid.


Did this page help you?