Wesloo
All topicsTopics

Cloud computing

Shared Responsibility Model in Cloud Security

Short answer

The shared responsibility model divides cloud security duties between the provider and the customer. The provider secures the underlying infrastructure, while the customer secures data, identities, applications, and configurations. The exact split changes across IaaS, PaaS, and SaaS.

Key points

  • Provider secures infrastructure, facilities, and the virtualization layer.
  • Customer owns data, identities, access controls, and endpoints.
  • IaaS leaves the most customer control; SaaS leaves the least.
  • Patch duties split: provider patches infrastructure, customer patches guest OS and apps.
  • Managed services shift some controls but not data or identity duties.

What is the shared responsibility model?

The shared responsibility model is the division of security duties between a cloud provider and its customer. The provider secures the underlying cloud infrastructure, while the customer secures the data, applications, identities, and configurations it places in that environment. The exact line between the two changes with the service model, so the same workload can leave different tasks to each party.

In an on-premises datacenter, one organization owns the whole stack. Moving to the cloud transfers some of those duties to the provider, but it does not transfer ownership of business data or user identities. For every deployment type, the customer still owns its data and identities and must protect them.

This guidance is usually framed as governance: it says who must set up, run, and watch each control. It is illustrative and does not create legal conclusions or change contract terms.

Security of the cloud and security in the cloud

A useful shorthand is security "of" the cloud compared with security "in" the cloud. The provider is responsible for the cloud itself: the hardware, software, networking, and facilities behind cloud services, along with the host operating system and virtualization layer. The customer is responsible for what it builds or runs inside that environment.

For example, in an infrastructure service the provider patches flaws in the underlying infrastructure, while the customer patches the guest operating system and applications. The customer also configures the provider's security group firewall. That split leaves physical security to the provider and workload configuration to the customer.

Customers should weigh the services they choose, how those services integrate with existing systems, and the laws and regulations that apply. The model is meant to provide flexibility and control, not a single fixed checklist.

How IaaS, PaaS, and SaaS change the boundary

The service model determines how much of the stack the customer manages. In infrastructure as a service, the customer manages virtual machines, operating systems, and applications, including Azure Disk Storage, Azure Virtual Machines, and virtual networks. In platform as a service, the customer deploys applications without managing virtual machines or operating systems, including Azure App Service, Azure SQL Database, Azure Functions, and Azure Storage. In software as a service, the customer uses ready-made applications such as Microsoft 365 and Dynamics 365.

Many real deployments mix these models, so responsibility can shift from one component to the next. A team might run a database on a managed platform while keeping custom code on virtual machines. For a broader comparison, see Differences Between IaaS, PaaS, and SaaS.

Serverless platforms are a common platform case: the customer writes functions and configures triggers, while the provider manages the execution environment. See What is serverless computing in Azure? for that model.

As the service model moves toward SaaS, the provider takes on more of the stack, but the customer keeps data, identity, and endpoint duties. In SaaS scenarios the provider may offer some device management capabilities, yet endpoint protection and compliance remain with the customer.

Duties that remain with the customer

Several duties do not move to the provider, even in managed services. The customer is responsible for its data, such as data classification, encryption decisions, data protection, and compliance with data governance requirements. It is also responsible for protecting endpoints and client devices used to access cloud services, such as mobile devices, laptops, and desktops.

Access management stays with the customer as well. That includes implementing and managing role-based access control, multifactor authentication, and conditional access policies. For managed services such as Amazon RDS and Amazon S3, the provider secures the service infrastructure and offers built-in security features, while the customer still manages access and data protection.

Application security is also a customer duty. That means controlling network exposure, safe application design, and regular reviews of security settings. Customers manage identities and permissions through proper IAM configuration so only authorized users and services can access resources.

Encryption decisions deserve explicit ownership. Customers protect sensitive data by setting encryption for stored and transmitted data, using provider tools according to their own security requirements. Leaving defaults in place without review is a common source of exposure.

What the provider secures

The provider secures the underlying cloud infrastructure. That includes physical data centers, server hardware, global networking components, and the facilities that house them. Providers apply strong physical and environmental security controls, including limiting access, monitoring facilities, and protecting hardware, to maintain infrastructure integrity, and they follow global compliance standards through recurring third-party audits.

For managed services, the provider secures the service infrastructure and may provide built-in security features. The provider also manages controls tied to the physical infrastructure that customers may have managed in their own datacenters. Because each customer deploys differently, this creates a distributed control environment rather than a single owner for every control.

Customers can use provider control and compliance documentation to perform their own control evaluation and verification procedures. That documentation matters when an organization must show auditors how a control is met in a shared environment.

Patching, controls, and compliance evidence

Patch management is a clear example of the split. The provider patches and fixes flaws within the infrastructure. The customer patches the guest operating system and applications. Missing either side leaves a gap, so patch ownership should be written into runbooks rather than assumed.

The shared model also extends to IT controls. Responsibility for operating the environment and for managing, operating, and verifying controls is shared. Shifting some control management to the provider can reduce operational burden, but the customer still needs to verify that its own controls work in the new environment.

Governance, compliance, and business continuity planning are affected by the same boundary. The service model and deployment model change who does which work, so a control that was internal in a private datacenter may become a shared or provider-managed control in the cloud.

Task ownership by service model

TaskIaaSPaaSSaaS
Physical facilities and hardwareProviderProviderProvider
Guest operating systemCustomerProviderProvider
Application codeCustomerCustomerProvider
Data and identitiesCustomerCustomerCustomer
Access policiesCustomerCustomerCustomer

This table is a simplification; mixed architectures and managed services can split a single task differently. It is still useful as a starting point for ownership reviews.

Why teams misread the model

A common mistake is to treat the provider's infrastructure security as complete workload security. The provider's duties stop at the infrastructure boundary; they do not include customer data classification, identity lifecycle, or endpoint hygiene. Another mistake is to assume that managed services remove all configuration work. Even when the provider secures the service infrastructure, the customer still manages access, data protection, and application settings.

Security groups, IAM policies, and encryption settings are customer-controlled in most services. If they are left at permissive defaults, the provider's physical security does not prevent misuse. Regular reviews of these settings catch drift before it becomes an incident.

A practical checklist for teams

A short review prevents the most common gaps:

  1. Map each workload to its service model and list who owns patching, configuration, and monitoring.
  2. Classify data and make encryption decisions for stored and transmitted information using provider tools according to your security requirements.
  3. Enforce least-privilege access with role-based access control, multifactor authentication, and conditional access.
  4. Protect endpoints that reach cloud services, such as mobile devices, laptops, and desktops.
  5. Review network exposure and application settings on a fixed schedule.
  6. Collect provider compliance documentation and match it to internal control tests.

For generative AI workloads and autonomous, tool-using AI agents, providers publish separate shared responsibility guidance. Those models add new questions about model inputs, tool permissions, and output handling, so they should be reviewed on their own terms.

Questions people also ask

Does the shared responsibility model make the provider liable for customer data breaches?

No. The model is governance guidance and does not change contract terms; customers remain responsible for their data and access controls.

Who patches what in the cloud?

The provider patches infrastructure flaws; the customer patches guest operating systems and applications.

Is security fully handled in SaaS?

No. Even in SaaS, customers still own data, identities, access policies, and endpoint protection.

Why does the boundary change by service model?

Each model leaves a different part of the stack under customer control, so duties move with the components the customer manages.

What evidence should customers collect?

Provider compliance documentation and internal control tests that show how shared controls are verified.

Related pages

  • Object vs block storage: key differences

    Object storage keeps data as self-contained objects in a flat namespace, while block storage splits data into fixed-size blocks that a file system or application addresses directly. The better choice depends on workload needs such as latency, throughput, scalability, and update patterns.

  • Cloud Virtualization: How It Works

    Cloud virtualization uses a hypervisor to split one physical server into multiple independent virtual machines. Providers allocate CPU, memory, storage, and network capacity to each guest, while users create and manage those machines through software interfaces. The result is higher hardware utilization for providers and flexible, on-demand scaling for users.

  • What is serverless computing in Azure?

    In serverless computing, the provider manages infrastructure and developers concentrate on code. In Azure, services such as Azure Functions and Azure Logic Apps support event-driven workflows, automatic scaling, and usage-based billing.

  • Differences Between IaaS, PaaS, and SaaS

    IaaS rents virtualized infrastructure, PaaS adds a managed development platform, and SaaS delivers finished applications. The main difference is how much of the stack the provider manages and how much control the customer keeps.

  • PID Controllers in Robotics: How Motion Control Works

    PID controllers in robotics compare a target motion with measured motion and combine proportional, integral, and derivative corrections to calculate a control command. The robot applies that command to a motor or other actuator, then repeats the measurement and correction as new sensor data arrives.

  • Electricity transmission over long distances

    Electricity transmission moves bulk power from generation sites to substations using high voltage, transformers, and interconnected lines. Raising voltage reduces current and resistive losses, while overhead and underground routes involve different costs, visibility, and repair trade-offs. Grid operators manage these networks, interconnections, and faults to keep power flowing across regions.

How this page was prepared

Before writing, the topic was researched in the sources below. The page then passed automatic quality checks.

Report a mistake on this page