Solutions Architect interview questions

Cloud computing provides technology resources on demand over a network with usage-based management.

• Resources include compute, storage, databases, networking, and managed application services.
• Customers can provision capacity through a console or API instead of buying and installing hardware first.
• Major providers such as AWS, Azure, and GCP operate the underlying data centers and service platforms.

IaaS, PaaS, and SaaS differ mainly in how much of the technology stack the provider manages.

• IaaS provides virtual machines, storage, and networking while the customer manages the operating system and application.
• PaaS provides a managed runtime and deployment platform while the customer supplies application code and data.
• SaaS provides a complete application used through a browser or API, such as Microsoft 365 or Salesforce.

The shared responsibility model divides security and operations between the cloud provider and the customer.

• The provider protects the physical facilities, hardware, and core managed-service infrastructure.
• The customer protects identities, data, configurations, and workloads according to the service being used.
• Customer responsibility is greater for IaaS and smaller for highly managed services such as SaaS.

A region is a geographic cloud location, while an availability zone is an isolated infrastructure location inside a region.

• A region is selected for latency, legal requirements, service availability, and business presence.
• Availability zones have separate power, networking, and facilities to limit correlated failures.
• Deploying across multiple zones can keep a regional application available when one zone fails.

These cloud models describe who operates the infrastructure and how an organization connects its environments.

• Public cloud uses shared provider infrastructure delivered as services to many customers with logical isolation.
• Private cloud is dedicated to one organization and may run in its own data center or with a hosting provider.
• Hybrid cloud connects private infrastructure with public cloud services so workloads and data can span both.

A cloud virtual machine is an isolated software-defined computer running on shared physical hardware.

• It has virtual CPU, memory, disks, networking, and its own guest operating system.
• A hypervisor separates multiple virtual machines on the same physical host.
• AWS EC2, Azure Virtual Machines, and GCP Compute Engine are common managed VM services.

Serverless computing runs code or applications without requiring the customer to provision or manage servers directly.

• The provider handles infrastructure, runtime availability, and much of the scaling.
• Function services such as AWS Lambda and Azure Functions execute code in response to requests or events.
• Billing commonly follows requests, execution time, or consumed resources rather than a permanently running server.

A container packages an application while sharing the host kernel, whereas a virtual machine includes a complete guest operating system.

• Containers are usually smaller and start faster because they do not boot a separate operating system.
• Virtual machines provide a stronger isolation boundary and can run different operating systems on one host.
• Docker builds and runs containers, while Kubernetes coordinates containers across a cluster.

A managed Kubernetes service operates the cluster control plane while customers deploy and manage their workloads.

• The provider maintains components such as the Kubernetes API and control-plane availability.
• Customers still manage workload definitions, container images, access policies, and often worker capacity.
• AWS EKS, Azure AKS, and Google GKE are the major cloud examples.

These AWS services cover common compute, container, database, and storage needs.

• EC2 provides virtual machines, Lambda runs event-driven functions, and EKS provides managed Kubernetes.
• RDS operates supported relational database engines such as PostgreSQL and MySQL.
• S3 stores objects such as files, backups, logs, and static website assets.

These Azure services provide three different ways to run application compute.

• Azure Virtual Machines provide operating-system-level control over persistent compute instances.
• Azure Functions runs event-driven code in a serverless execution model.
• AKS provides managed Kubernetes for orchestrating containerized applications.

These GCP services run workloads at virtual-machine, managed-container, and Kubernetes levels.

• Compute Engine provides configurable virtual machines.
• Cloud Run runs container images on a managed platform and scales instances according to requests.
• GKE provides managed Kubernetes clusters for container orchestration.

Object, block, and file storage organize and access data through different interfaces.

• Object storage keeps complete objects with metadata in buckets and is accessed through an API.
• Block storage exposes raw volumes that an operating system formats and mounts like a disk.
• File storage presents shared directories and files through protocols such as NFS or SMB.

Amazon S3 is an object storage service that stores objects inside buckets.

• Each object has a key, content, metadata, and optional version information.
• Buckets have names, regional placement, access policies, and configuration such as lifecycle rules.
• Applications access S3 through APIs rather than mounting it as a normal block disk.

Durability measures whether stored data remains intact, while availability measures whether the service can be used now.

• A durable system keeps data safe despite hardware failures through replication or error correction.
• An available system accepts and serves requests during the measured period.
• A storage service can temporarily be unavailable without losing data, so the two metrics are not interchangeable.

Versioning preserves multiple object revisions, while lifecycle policies automate how objects are retained or moved.

• Versioning can protect against accidental overwrite or deletion by keeping older versions.
• Lifecycle rules can move old objects to cheaper storage classes or delete them after a defined period.
• Both features require retention and cost planning because preserved versions still consume storage.

A VPC or virtual network is a logically isolated network space inside a cloud provider.

• It has an IP address range divided into subnets for organizing resources.
• Route tables, gateways, and security controls determine where traffic can flow.
• AWS calls it a VPC, while Azure commonly calls it a Virtual Network or VNet.

A public subnet has a route for direct internet connectivity, while a private subnet does not expose resources directly that way.

• Public subnets commonly contain internet-facing load balancers or gateways.
• Private subnets commonly contain application servers and databases reached through internal routes.
• A resource also needs suitable addressing and security rules, so subnet type alone does not guarantee access.

A route table maps destination network ranges to the next hop for traffic.

• A local route allows communication among subnets inside the virtual network.
• A default route can send internet-bound traffic to an internet or NAT gateway.
• The most specific matching route normally determines which path a packet takes.

Cloud firewall rules and security groups control allowed network traffic to and from resources.

• Rules typically match protocol, port, source, and destination.
• Inbound rules govern traffic entering a resource, while outbound rules govern traffic leaving it.
• Least-privilege rules allow only the paths required by the application instead of broad public access.