The AWS Shared Responsibility Model: Understanding Your Security Responsibilities
Moving an application to AWS does not mean that security becomes entirely AWS's responsibility. One of the most important concepts to understand when designing secure cloud environments is the AWS Shared Responsibility Model. The model defines how security responsibilities are divided between AWS and the customer. AWS is responsible for security of the cloud, while customers are responsible for security in the cloud. Exactly where that boundary falls depends on the AWS services being used.
The AWS Shared Responsibility Model: Understanding Your Security Responsibilities
Understanding this distinction is important because a secure AWS infrastructure can still contain vulnerabilities caused by incorrect customer configuration.
Security of the Cloud
AWS is responsible for protecting the infrastructure that runs AWS services.
This includes the underlying:
- Physical facilities
- Hardware
- Networking infrastructure
- Storage infrastructure
- Compute infrastructure
- Virtualization layer
- AWS global infrastructure
AWS manages security controls at these infrastructure layers and continuously operates the infrastructure supporting its cloud services.
For example, when an organization deploys an Amazon EC2 instance, the physical server hosting that instance is part of the infrastructure protected and managed by AWS.
However, AWS securing the physical server does not automatically secure everything running on the EC2 instance.
That is where the customer's responsibility begins.
Security in the Cloud
Customers are responsible for securing the resources, applications, configurations, identities, and data they deploy in AWS.
The exact responsibilities depend on the service.
For example, with Amazon EC2, customers are responsible for:
- The guest operating system
- Operating system security patches
- Applications installed on the instance
- Security group configuration
- Access permissions
- Data stored on the instance
AWS provides the infrastructure, but the customer must securely configure and operate the workload running on it.
This creates an important security principle:
Using a secure cloud provider does not automatically create a secure cloud environment.
The Responsibility Changes Depending on the Service
One of the most useful aspects of the Shared Responsibility Model is that the customer's security responsibilities can change depending on how much of the underlying infrastructure AWS manages.
Consider three examples: EC2, Amazon S3, and AWS Lambda.
Amazon EC2
EC2 provides virtual servers while giving customers significant control.
The customer manages the guest operating system, installed software, applications, security groups, and other configurations.
For example, if an EC2 instance is running an outdated Linux package containing a vulnerability, patching that operating system is the customer's responsibility.
Amazon S3
With Amazon S3, AWS manages much more of the underlying infrastructure and platform.
The customer's focus shifts toward protecting the data and controlling access.
Important customer responsibilities include:
- Identity and access management
- Bucket and object permissions
- Data protection
- Encryption configuration
- Data classification
- Appropriate access policies
Therefore, an accidentally public S3 bucket is generally a customer configuration problem rather than an AWS infrastructure problem. AWS provides the security controls, but customers must configure them appropriately.
AWS Lambda
Lambda moves even more infrastructure management to AWS.
AWS manages the underlying infrastructure, operating system, execution environment, and runtime platform. Customers remain responsible for their function code, configuration, and IAM permissions.
This illustrates an important cloud-security concept:
The more infrastructure AWS manages, the less infrastructure the customer needs to secure directly—but the customer's responsibility for identities, configurations, applications, and data does not disappear.
A Simple Way to Think About the Model
A useful mental model is to divide responsibilities into two layers.
AWS
Infrastructure → Hardware → Networking → Physical facilities → Virtualization
Customer
Data → Applications → Identities → Permissions → Configuration → Workload security
The exact boundary can move depending on the service being used.
For example:
Area | EC2 | S3 | Lambda |
|---|---|---|---|
Physical infrastructure | AWS | AWS | AWS |
Hardware | AWS | AWS | AWS |
Operating system | Customer | AWS | AWS |
Application/code | Customer | Customer | Customer |
Identity & permissions | Customer | Customer | Customer |
Data protection | Customer | Customer | Customer |
Service configuration | Customer | Customer | Customer |
The table is a simplified representation. AWS responsibilities vary by service and configuration, so the specific AWS documentation for each service should always be consulted.
Common Security Mistakes
Understanding the Shared Responsibility Model also helps explain why many cloud security incidents involve configuration or access-control issues.
1. Assuming AWS patches everything
AWS is responsible for patching vulnerabilities in the infrastructure it manages.
That does not mean AWS automatically patches the operating system and applications running inside your EC2 instances.
For EC2, operating-system and application patching remain customer responsibilities.
2. Giving excessive IAM permissions
AWS provides IAM to control authentication and authorization, but customers must decide what permissions users, roles, and applications receive.
A better approach is least privilege: give an identity only the permissions required to perform its task.
For example, an application that only needs to read objects from one S3 bucket should not automatically receive administrator permissions across the AWS account.
3. Treating encryption as entirely AWS's responsibility
AWS provides encryption capabilities across many services, but customers are responsible for determining how their data should be protected and configuring the appropriate controls.
Security requirements should therefore be considered when designing the workload rather than after the application has already been deployed.
4. Ignoring logging and monitoring
A secure environment also needs visibility.
Customers should understand what activities are occurring within their AWS environments and use appropriate AWS security and monitoring services to identify suspicious activity and configuration problems.
Security is therefore not just about preventing attacks. It also involves detection, investigation, and response.
Applying the Model to a Real Application
Consider a simple web application running on AWS:
Internet → Load Balancer → EC2 → Application → S3 Database/Data
The responsibilities might look like this:
AWS manages:
- Physical data centers
- Physical servers
- Underlying networking
- Virtualization infrastructure
The customer manages:
- EC2 operating system
- Security groups
- Application security
- IAM permissions
- S3 access policies
- Data encryption configuration
- Application secrets
- Logging and monitoring configuration
This means that even if AWS infrastructure is properly secured, the application could still be vulnerable because of a weak IAM policy, an exposed secret, an incorrectly configured security group, or an unpatched operating system.
The Shared Responsibility Model helps identify where each control belongs.
How Security Teams Can Use the Model
The model can also be used as a practical checklist when designing or reviewing an AWS environment.
For each AWS service, ask:
- What does AWS manage?
- What do we manage?
- What data are we storing?
- Who can access that data?
- Which identities and roles have permissions?
- Which systems require patching?
- Which security configurations must we maintain?
- How are security events logged and monitored?
- What security requirements apply to the workload?
This approach prevents teams from assuming that a security control exists simply because the workload is hosted on AWS.
AWS also recommends reviewing the security functionality and configuration options of individual services, considering organizational and regulatory requirements, and evaluating AWS security services against those requirements.
Conclusion
The AWS Shared Responsibility Model is one of the foundations of AWS security.
AWS is responsible for security of the cloud, including the infrastructure that operates AWS services. Customers are responsible for security in the cloud, including the configuration, identities, applications, and data associated with their workloads.
The most important lesson is that moving to AWS does not remove the customer's security responsibilities. Instead, it changes them.
The right question is not:
"Is AWS secure?"
The better question is:
"Which parts of my environment does AWS secure, and which parts do I need to secure myself?"
Once that distinction is understood, security teams and developers can design AWS workloads with clearer ownership, stronger access controls, better monitoring, and fewer configuration-related security risks.
References
- AWS Shared Responsibility Model
- AWS Well-Architected Framework — Security Pillar
- AWS Risk and Compliance Whitepaper
- AWS Lambda Security Overview
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article