Skip to content
Tutorials Center
My learning

Topics

Lesson 4 of 70 ยท Getting started with AWS

The Shared Responsibility Model

Understand what AWS secures and what you secure for EC2, RDS, S3 and Lambda, how compliance reports fit in, and the six Well-Architected pillars, with a runnable checklist classifier.

When you run software on AWS, security and reliability are split between two parties. AWS protects the cloud itself: buildings, hardware, the virtualization layer and the managed service software. You protect what you put in the cloud: your data, your configuration, your access rules and, depending on the service, your operating system and code. This split is called the shared responsibility model, and misunderstanding it is behind most cloud security incidents. Almost every publicized "AWS breach" turns out to be a customer-side misconfiguration, such as a public bucket or an exposed access key, not a failure of AWS's data centers.

This lesson gives you a way to answer, for any service, "who handles this?" It also introduces the Well-Architected Framework, the checklist AWS publishes for judging a design. The two runnable examples are stand-ins: a classifier that assigns tasks to AWS or to you for four services, and a small review tool that groups questions by pillar. Real responsibility depends on the exact service and settings, so treat these as practice material, not as an official matrix.

Security of the cloud and security in the cloud

AWS describes the split in two phrases:

  • Security of the cloud is AWS's job: physical security of the data centers, power and cooling, the network backbone, the hosts and hypervisors, and the software of the managed services.
  • Security in the cloud is your job: who may access your account, how data is classified and encrypted, how networks and firewalls are configured, and the software you install.
flowchart TB
    subgraph You["You: security IN the cloud"]
        D["Data and encryption choices"]
        I["Identity and access"]
        C["Configuration and firewall rules"]
    end
    subgraph AWS["AWS: security OF the cloud"]
        S["Managed service software"]
        V["Virtualization and hosts"]
        H["Hardware, network, data centers"]
    end
    You --> AWS

Your responsibilities sit on top of AWS's, and the boundary moves with the service you choose.

The boundary is not fixed. The more managed the service, the more AWS takes over. That is the connection to IaaS, PaaS and FaaS from What is AWS?.

Examples by service

Here is how the responsibility shifts for four services you will meet in this tutorial.

Amazon EC2 (virtual machines). AWS secures the hardware and the hypervisor. You choose the operating system image, and you must patch it, harden it, install antivirus if you need it, configure the security groups, manage the users on the machine and protect the data on the disks. If a server is unpatched, that is on you.

Amazon RDS (managed databases). AWS patches the operating system and offers engine patching during maintenance windows, handles backups and replaces failed hardware. You control who can connect (security groups, database users and passwords), whether the data is encrypted, how long backups are kept and the schema and queries you run.

Amazon S3 (object storage). AWS keeps the storage infrastructure running and stores your data redundantly. You decide who can read or write each bucket through policies, whether public access is blocked, whether objects are encrypted with a key you control, and whether versioning is on. Almost every S3 data leak comes from the access decisions in your column.

AWS Lambda (functions). AWS manages the servers, scaling and patching of the managed runtime. You write the code, choose and update the libraries it depends on, give the function an execution role with only the permissions it needs, and protect the data it handles. You do not patch an operating system, but you do own vulnerabilities in your code and dependencies.

Notice the pattern: identity, data and configuration are always yours, on every service.

Shared controls

A few controls are shared between both parties, each side doing its own part:

  • Patch management: AWS patches the infrastructure and managed service software; you patch guest operating systems and applications.
  • Configuration management: AWS configures its devices; you configure your own resources, guest operating systems, databases and applications.
  • Awareness and training: AWS trains its employees; you train yours.

Try it: who owns this task?

The classifier below holds a simple table of tasks and the owner for each of four services. It is a teaching simplification (reality has gray areas, such as engine patch scheduling on RDS, where AWS runs the patch in a window you choose), but the pattern is accurate.

Python Assign each task to AWS or you

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Run it and compare the totals for EC2 and Lambda. Then add a service of your own (for example "sqs") by adding a key to each row, and decide who owns each task.

Compliance in a shared model

Many industries need proof that systems meet standards such as ISO 27001, SOC 2 or PCI DSS. AWS has audits of its own part of the model and publishes the reports through AWS Artifact, a console page where you download audit reports and accept agreements. These reports show that AWS's data centers and services meet the standards. They do not make your application compliant. You still have to configure your part correctly, and you may need to show your own auditor the evidence: access logs, encryption settings, patch records. Later chapters give you the tools for that, including CloudTrail, Config and X-Ray.

Data location is part of your responsibility too. Data stays in the Region where you put it unless you move or replicate it, so choosing a Region is also a legal decision.

The Well-Architected pillars

AWS publishes a framework for evaluating designs, called the Well-Architected Framework. It organizes questions under six pillars:

  1. Operational excellence: run and monitor systems, and improve procedures over time.
  2. Security: protect data, systems and assets.
  3. Reliability: recover from failures and meet demand.
  4. Performance efficiency: use resources efficiently as needs change.
  5. Cost optimization: avoid unnecessary spending.
  6. Sustainability: minimize the environmental impact of workloads.

Pillars pull against each other: the most reliable design is usually more expensive, and the cheapest one often has fewer safeguards. The framework does not choose for you; it makes you decide on purpose. You will study this properly in Well-Architected Framework. For now, the stand-in below shows the review habit: ask questions per pillar and count the gaps.

Python A tiny pillar review

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Check your understanding

Quick check ยท 5 questions

of 5 answered You scored / 5 ยท perfect!

  1. 1. Who is responsible for patching the guest operating system of an EC2 instance?

    EC2 is IaaS. AWS secures the hardware and hypervisor; you manage the operating system and everything above it.
  2. 2. An S3 bucket exposes customer files to the whole internet. Whose responsibility is the configuration that allowed it?

    AWS runs the storage service, but who may read a bucket is a configuration decision that belongs to the customer.
  3. 3. What does an AWS audit report from AWS Artifact prove?

    The reports cover AWS's part of the model. You must still configure and document your own part.
  4. 4. Which of these is one of the six Well-Architected pillars?

    The pillars are operational excellence, security, reliability, performance efficiency, cost optimization and sustainability.
  5. 5. Which responsibility is yours on every AWS service?

    Who can access your data and how it is protected is always a customer decision, even for the most managed services.

Summary

  • AWS secures the cloud itself; you secure what you put in it.
  • The more managed the service (EC2, then RDS, then Lambda), the less you operate, but identity, data and configuration are always yours.
  • Patching, configuration and training are shared controls with a part for each side.
  • AWS Artifact gives you AWS's audit reports; they do not make your own application compliant.
  • The Well-Architected Framework has six pillars that you trade off on purpose.

Next: IAM Users, Groups and Roles.