Home Blog Confidential Computing and P Azure confidential VMs: DCsv3 deployment guide
🔒 Confidential Computing and P May 22, 2026 12 min read

Azure confidential VMs: DCsv3 deployment guide

Confidential Computing and P Enterprise Guide 2026 SCALE D2C D2C Technology Confidential Computing and P Enterprise Guide 2026 SCALE D2C D2C Technology

Azure Confidential VMs — the DCsv3, DCdsv3, and DCasv5 series — bring Intel SGX and AMD SEV-SNP hardware-based confidential computing to enterprise Azure workloads. They are the practical on-ramp to confidential computing for Azure enterprises: standard VMs with hardware-enforced isolation, requiring minimal application changes. This deployment guide covers DCsv3 vs DCasv5 selection, attestation setup, performance characteristics, and enterprise integration patterns.

Azure Confidential VM Series Overview

SeriesTechnologyvCPU rangeMemoryBest For
DCsv3 / DCdsv3Intel SGX (application enclave)1–96 vCPUUp to 384 GiBApplication-level enclaves — specific sensitive workloads
DCasv5 / DCadsv5AMD SEV-SNP (full VM encryption)2–96 vCPUUp to 384 GiBLift-and-shift confidential VMs — entire VM protected
ECasv5AMD SEV-SNP (memory-optimised)2–96 vCPUUp to 672 GiBMemory-intensive confidential workloads — databases, analytics
DCsv3 vs DCasv5: The Key Decision
DCsv3 (Intel SGX): Requires application-level code changes to use the SGX enclave APIs — but provides the strongest isolation for specific sensitive operations (key processing, cryptographic operations). Higher engineering effort; strongest protection for specific operations.
DCasv5 (AMD SEV-SNP): Protects the entire VM with hardware memory encryption — no application code changes required. Lift-and-shift your existing workload to a confidential VM. For most enterprise use cases requiring data-in-use protection without code refactoring, DCasv5 is the right starting point.

DCsv3 Deployment Guide

01
Step 1
Provision DCsv3 VM

Deploy DCsv3 from Azure portal or via Terraform. Select Ubuntu 22.04 or Windows Server 2022 — both include Intel SGX SDK support. Verify SGX availability: ls /dev/sgx_enclave (should exist). Install Azure DCAP (Data Center Attestation Primitives) library: sudo apt-get install az-dcap-client. This library enables communication with Intel's PCCS attestation service via Azure's infrastructure — required for remote attestation. Provision via your existing infrastructure-as-code tooling.

Terraform provisioningaz-dcap-clientSGX device verification
02
Step 2
Build and Deploy Enclave Application

Use Open Enclave SDK (cross-platform, recommended) or Intel SGX SDK for enclave development. Structure your application: trusted component (enclave code — processes sensitive data), untrusted host (normal process — handles I/O, calls enclave via ECALL/OCALL). Build enclave as a signed .so shared library. Deploy in containerised form using EGo (Go in SGX enclaves) or Gramine (run existing Linux applications in SGX with minimal modifications). Integrate with your CI/CD pipeline.

Open Enclave SDKGramine for lift-and-shiftContainer deployment
03
Step 3
Remote Attestation with Microsoft Azure Attestation

Use Microsoft Azure Attestation (MAA) service — Azure-native attestation for both SGX and SEV-SNP. MAA verifies the enclave quote and issues a JWT attestation token that relying parties (key management systems, other services) can verify. Integrate attestation into your application's startup flow: verify MAA token before releasing sensitive data to the enclave. Connect Azure Key Vault with attestation conditions — keys only released to verified enclaves.

Microsoft Azure AttestationJWT attestation tokenAzure Key Vault integration
0
Code changes required to use DCasv5 AMD SEV-SNP VMs — lift your existing workload into a fully hardware-encrypted VM. The most practical on-ramp to Azure confidential computing for most enterprises
5–10%
Performance overhead for DCasv5 SEV-SNP VMs vs equivalent non-confidential VMs — acceptable for most enterprise workloads requiring data-in-use protection
All
Three major cloud providers now offer confidential VM instances — AWS Nitro Enclaves, Azure DCasv5/DCsv3, Google Confidential C3 — making confidential computing accessible at standard cloud pricing
🏥
Healthcare AI on Azure
Deploy PHI-processing AI workloads on DCasv5 — the entire VM is hardware-encrypted, satisfying HIPAA data-in-use requirements without application code changes. Combine with Azure HIPAA/HITRUST Blueprint for a complete compliant architecture. Our healthcare app development team deploys confidential VM architectures for health system clients.
💳
Financial Cryptography
Run HSM-equivalent cryptographic operations in DCsv3 SGX enclaves — private keys processed inside the enclave are never exposed to the OS, hypervisor, or Azure operators. For payment processors, custodians, and certificate authorities requiring cloud-scale HSM functionality, DCsv3 provides the security model at a fraction of dedicated HSM hardware cost.
Azure Confidential VM Deployment

Our software development and DevOps teams design and deploy Azure confidential VM architectures for regulated enterprise workloads. Book a free advisory session to scope your confidential computing architecture on Azure.

Frequently Asked Questions

End-to-end Confidential Computing and P strategy, implementation, and optimisation for enterprise and D2C brands. Contact us for a free consultation.

Strategy projects: 4–8 weeks. Full implementation: 3–12 months. ROI typically within 12–18 months.

Yes — D2C brands to enterprise. View our pricing.

CONFIDENTIAL

Ready to Implement Confidential Computing and P?

Our specialist team delivers measurable ROI from Confidential Computing and P programmes for enterprise and D2C brands.

Free Audit