Architecture models
Exam objective 3.1: Compare and contrast security implications of different architecture models.
On this page
Cloud, on-premises, hybrid, serverless, microservices, IoT, and industrial control systems each move risk around. This objective asks what you gain and what you take on with each model, including the shared responsibility model, infrastructure as code, and considerations like availability, cost, and patching.
Cloud and responsibility
In the cloud, security is split between provider and customer. The provider always secures the physical facilities and hardware. As you move from infrastructure as a service (IaaS) to platform (PaaS) to software (SaaS), the provider takes on more: the OS, then the runtime, then the application. The customer always owns its data, its identities, and how it configures access.
| Layer | IaaS | PaaS | SaaS |
|---|---|---|---|
| Data and access | Customer | Customer | Customer |
| Application | Customer | Customer | Provider |
| OS and runtime | Customer | Provider | Provider |
| Hardware and facility | Provider | Provider | Provider |
Hybrid architectures connect on-premises and cloud, which adds the burden of consistent identity and policy across both. Third-party vendors in the chain need contractual security requirements.
Modern application models
Infrastructure as code (IaC) defines servers, networks, and permissions in version-controlled templates, so environments are repeatable and reviewable; a mistake in a template is also repeated everywhere. Serverless functions remove OS management but shift focus to code, permissions, and event inputs. Microservices split an application into small services that talk over APIs, which improves isolation but multiplies the number of interfaces to secure. Containers package an app with its dependencies and share the host kernel, so a kernel flaw affects every container on it.
Network infrastructure can be physically isolated (air-gapped), logically segmented, or software-defined networking (SDN), where a controller programs the network centrally. Centralized designs are easier to manage and monitor but create single points of failure; decentralized designs are more resilient but harder to keep consistent.
Embedded, IoT, and industrial systems
Industrial control systems (ICS) and SCADA run factories, utilities, and pipelines. They value availability above all, often can't be patched quickly, and use old protocols. Real-time operating systems (RTOS) power devices that must respond within strict time limits. Embedded systems and IoT devices are often low-power with limited compute, so they may lack encryption or update mechanisms.
| Consideration | Question to ask |
|---|---|
| Availability and resilience | What happens when one part fails? |
| Cost | Up front and ongoing, including staff |
| Responsiveness and scalability | Can it handle peaks? |
| Ease of deployment and recovery | How fast can we rebuild it? |
| Patch availability | Can we patch, and how fast? |
| Power and compute | Can the device run security controls at all? |
| Risk transference | Can a provider or insurer take on part of the risk? |
Key terms
- Shared responsibility model
- The division of security duties between a cloud provider and its customer, which shifts with the service model.
- Infrastructure as a service (IaaS)
- A cloud model where the provider supplies virtual hardware and the customer manages the OS and everything above it.
- Platform as a service (PaaS)
- A cloud model where the provider manages the OS and runtime and the customer deploys its own code.
- Software as a service (SaaS)
- A cloud model where the provider runs the whole application and the customer manages data and user access.
- Hybrid cloud
- An architecture that combines on-premises infrastructure with public cloud services.
- Infrastructure as code (IaC)
- Defining and provisioning infrastructure with version-controlled templates instead of manual setup.
- Serverless computing
- Running code as functions on demand without managing servers, while the provider handles the OS.
- Microservices
- An application design that splits functions into small independent services communicating over APIs.
- Air gap
- Physical separation of a network or system from all other networks, including the internet.
- Software-defined networking (SDN)
- A design where a central controller programs network behavior separately from the hardware that forwards traffic.
- Industrial control system (ICS) / SCADA
- Systems that monitor and control physical industrial processes such as power, water, and manufacturing.
- Real-time operating system (RTOS)
- An operating system that guarantees processing within strict time limits for embedded devices.
- Internet of Things (IoT)
- Network-connected everyday devices, often with limited compute, power, and update support.
- Containerization
- Packaging an application with its dependencies to run in isolated user spaces that share the host kernel.
Check yourself
28 practice questions on this objective
Practice this objectiveFlashcards for this objective