AWS vs Microsoft Azure

What is AWS?

Amazon Web Services launched in 2006 and was the first hyperscaler to offer infrastructure as a service at scale. AWS has the broadest service catalogue — over 200 services spanning compute (EC2, Lambda), storage (S3, EBS), databases (RDS, DynamoDB, Redshift), machine learning (SageMaker), networking, and developer tools. Its identity model (IAM) is granular and policy-driven, and its global region footprint is unmatched.

AWS's strength is breadth and maturity. Teams building cloud-native workloads on Linux, using open-source databases, or deploying microservices architectures often favour AWS for the depth of available tooling and the size of the community. AWS also leads in serverless innovation — Lambda's execution model, Fargate for containerised workloads, and Step Functions for orchestration have broad enterprise adoption.

What is Microsoft Azure?

Azure launched in 2010 and was initially positioned around Windows Server and .NET. Today it is a full hyperscaler, but its differentiation remains the Microsoft ecosystem: Active Directory, Office 365, Teams, SQL Server, Windows Server, and Visual Studio all integrate natively with Azure services.

For enterprises already running Microsoft infrastructure, Azure provides a coherent platform: Entra ID (formerly Azure Active Directory) extends on-premises AD to the cloud, Azure SQL Managed Instance offers near-100% SQL Server compatibility, and Azure DevOps provides an integrated CI/CD and project management suite. Azure's hybrid story — led by Azure Arc — is widely considered the most mature in the industry, allowing organisations to project Azure governance and management onto on-premises Kubernetes clusters, bare-metal servers, and even workloads running on AWS or GCP.

Enterprise integration: Active Directory, Office 365, and .NET

The single biggest reason enterprises choose Azure over AWS is identity. Microsoft Entra ID is the cloud extension of Active Directory — it provides single sign-on (SSO) for thousands of SaaS applications, multi-factor authentication, conditional access policies, and Privileged Identity Management. When a user signs into Office 365 or Teams, they are authenticating via Entra ID. Extending this to cloud workloads on Azure requires no additional identity infrastructure.

On AWS, federating with Active Directory is possible but requires AWS IAM Identity Centre, Active Directory Connector, or a third-party identity provider. It works, but there are more moving parts.

.NET Framework applications — still common in large enterprises — run on Azure App Service without modification. Azure supports the full .NET ecosystem: .NET Framework (for legacy apps), .NET 6/8 (modern), and Blazor (browser-side C#). Azure Functions natively support C# in-process and isolated worker models. For teams maintaining large Windows-based line-of-business applications, Azure reduces migration friction significantly.

Hybrid cloud: Azure Arc vs AWS Outposts

Both AWS and Azure offer hybrid cloud solutions, but their approaches differ fundamentally.

AWS Outposts delivers physical AWS-managed hardware racks to your on-premises data centre. The hardware runs the same AWS services (EC2, EKS, RDS) and connects back to the AWS region. This is excellent if you want identical AWS APIs on-premises, but it requires physical delivery, installation, and a minimum size commitment — and it only covers AWS.

Azure Arc is software-only. It projects the Azure control plane — Azure Resource Manager, Policy, RBAC, Defender for Cloud, Monitor — onto existing infrastructure regardless of where it runs: on-premises servers, Kubernetes clusters on AWS or GCP, or edge devices. You manage everything from the Azure portal using the same tooling. For brownfield enterprises with mixed estates, this is a significant advantage. Azure Arc-enabled data services can also run Azure SQL and PostgreSQL on-premises with cloud-managed updates.

DevOps pipelines and CI/CD

Azure DevOps is a comprehensive platform covering the full software delivery lifecycle: Boards (work items, sprints, Kanban), Repos (Git hosting), Pipelines (CI/CD), Artifacts (package registry), and Test Plans (manual and automated testing). Many enterprises use Azure DevOps as their primary engineering platform regardless of which cloud they deploy to.

AWS has CodePipeline, CodeBuild, CodeDeploy, and CodeCommit, but these are often used in combination with GitHub Actions, Jenkins, or CircleCI rather than as a standalone suite. AWS CodeStar (now deprecated) was an attempt at an integrated experience, but the ecosystem never cohered the same way. If your team already uses Azure DevOps, staying on Azure for deployment is the path of least resistance. For teams building Kubernetes-based deployments or using GitOps patterns, both clouds support Flux and ArgoCD equally well.

Compliance and government cloud

Both AWS and Azure hold an extensive range of compliance certifications: ISO 27001, SOC 1/2/3, PCI DSS, HIPAA, Cyber Essentials Plus (UK), and many others. The right choice depends on your regulatory framework and geography.

Azure has a notable advantage in UK and European regulated sectors. Microsoft was early to establish UK-specific regions (UK South, UK West) and has long-standing relationships with UK government via the Crown Commercial Service (CCS) framework. Azure Government (US) and Azure for Sovereign Clouds (EU) provide dedicated environments for government workloads. Microsoft Purview (compliance management) and Microsoft Defender for Cloud are deeply integrated.

AWS GovCloud (US-East and US-West) serves US federal agencies and contractors. AWS also holds FedRAMP High authorisation across a broad service set. For global enterprises, AWS's larger region footprint can be an advantage for data residency requirements.

Pricing and cost models

Pricing comparisons between AWS and Azure are notoriously unreliable at a general level because they depend heavily on workload type, OS, region, and contract terms. However, a few patterns are well established:

• Windows workloads: Azure Hybrid Benefit can reduce Windows Server VM costs by up to 40% and SQL Server costs by up to 85% for organisations with Software Assurance. AWS has no equivalent for Microsoft licences.
• Linux compute: AWS tends to be competitive or cheaper, with a wider range of instance families (Graviton ARM instances offer strong price-performance).
• Reserved capacity: Both offer 1-year and 3-year reservations. AWS Reserved Instances and Savings Plans are more flexible; Azure Reserved VM Instances are straightforward.
• Egress costs: Both charge for outbound data transfer, which can be a significant cost driver. Negotiate egress commitments as part of enterprise agreements.

Below is an example Azure CLI command to get pricing information for a specific VM size:

# Query Azure retail prices for a specific VM size in UK South
az rest --method get \
  --url "https://prices.azure.com/api/retail/prices?\$filter=armRegionName eq 'uksouth' and skuName eq 'D4s v5'" \
  --query "Items[?type=='Consumption'].{SKU:skuName,Price:retailPrice,Unit:unitOfMeasure}"

And an AWS equivalent using the CLI to retrieve on-demand pricing:

# Get on-demand price for m5.xlarge in eu-west-2 (London)
aws pricing get-products \
  --service-code AmazonEC2 \
  --filters 'Type=TERM_MATCH,Field=instanceType,Value=m5.xlarge' \
             'Type=TERM_MATCH,Field=location,Value="EU (London)"' \
             'Type=TERM_MATCH,Field=operatingSystem,Value=Linux' \
  --region us-east-1 \
  --query 'PriceList[0]' | python3 -m json.tool | grep -A2 '"USD"'