Cloud Solutions Architect Interview Flashcards

Question 1

What is cloud computing?

Accepted Answer

Cloud computing delivers compute, storage, databases, and networking on demand over the internet, billed by usage rather than owned upfront. It removes the need to buy and maintain hardware, enabling elastic scaling, faster delivery, and a shift from capital to operating expense.

Question 2

What is the difference between IaaS

Accepted Answer

IaaS provides virtualized infrastructure (VMs, storage, networking) you manage, like EC2. PaaS provides a managed runtime where you deploy code without managing servers, like App Engine. SaaS delivers complete applications over the web, like Gmail. Moving up the stack trades control for less operational burden.

Question 3

What is the difference between public

Accepted Answer

Public cloud uses a provider's shared, multi-tenant infrastructure (AWS, Azure, GCP). Private cloud is dedicated to one organization for control or compliance. Hybrid cloud connects on-premises/private with public cloud, letting workloads run where they best fit for cost, latency, or regulation.

Question 4

What is elasticity in the cloud?

Accepted Answer

Elasticity is the ability to automatically scale resources up during demand spikes and down during lulls, so you pay only for what you need. It differs from raw scalability by being dynamic and automated, enabling cost efficiency and consistent performance under variable load.

Question 5

What is the difference between vertical and horizontal scaling?

Accepted Answer

Vertical scaling (scale up) increases a single instance's CPU, memory, or storage, simple but capped by hardware limits and a single point of failure. Horizontal scaling (scale out) adds more instances behind a load balancer, offering near-unlimited growth and resilience but requiring stateless, distributable design.

Question 6

What is a cloud region?

Accepted Answer

A region is a distinct geographic location containing multiple isolated data centers (availability zones). Choosing a region affects latency to users, data residency and compliance, available services, and cost. Workloads are often deployed to regions near users or required by regulation.

Question 7

What is an availability zone?

Accepted Answer

An availability zone (AZ) is one or more physically separate data centers within a region with independent power, cooling, and networking. Distributing resources across AZs protects against a single data center failure, forming the basis of high-availability architectures within a region.

Question 8

What is a virtual machine in the cloud?

Accepted Answer

A virtual machine is an isolated, software-defined computer with its own OS running on a hypervisor that shares physical hardware among tenants. It gives full control over the environment (IaaS), is quick to provision, and can be resized, snapshotted, and replaced on demand.

Question 9

How does a container differ from a virtual machine?

Accepted Answer

A container packages an application with its dependencies and shares the host operating system kernel, making it lightweight, fast to start, and portable. A VM virtualizes hardware and runs a full guest OS, giving stronger isolation but more overhead. Containers enable dense, consistent deployments via images.

Question 10

What does a load balancer do?

Accepted Answer

A load balancer spreads incoming requests across multiple backend instances to improve availability and throughput, performs health checks to route around failed instances, and can terminate TLS. Layer 4 balances by IP/port; layer 7 routes by HTTP attributes like path or host.

Question 11

What is auto scaling?

Accepted Answer

Auto scaling automatically adds or removes compute instances according to metrics (CPU, request count) or schedules, maintaining performance during spikes and cutting cost during lulls. It pairs with a load balancer and requires stateless instances so any can serve any request.

Question 12

What is object storage used for?

Accepted Answer

Object storage (like S3) stores data as objects with metadata in a flat namespace accessed via API, scaling massively and cheaply with high durability. It suits unstructured data, backups, static assets, and data lakes, but is not a filesystem and has higher latency than block storage.

Question 13

What is block storage best suited for?

Accepted Answer

Block storage presents raw volumes (like EBS) attached to a single instance, offering low latency and high IOPS for databases, boot disks, and transactional workloads. Unlike object storage it behaves as a disk you format and mount, and is typically tied to one instance at a time.

Question 14

What does a CDN do?

Accepted Answer

A CDN caches static and sometimes dynamic content at globally distributed edge locations near users, reducing latency, offloading origin servers, and improving availability. It speeds delivery of images, video, and assets, and can absorb traffic spikes and mitigate some DDoS attacks.

Question 15

What is a VPC?

Accepted Answer

A VPC is a logically isolated virtual network in the cloud where you define IP ranges, subnets, route tables, and gateways, controlling connectivity and security. It isolates your resources from other tenants and lets you connect securely to on-premises networks via VPN or direct links.

Question 16

What is a subnet in a VPC?

Accepted Answer

A subnet subdivides a VPC's IP range, usually scoped to one availability zone. Public subnets route to an internet gateway for internet-facing resources; private subnets have no direct internet route, hosting databases and internal services that reach out via NAT. Subnetting structures security and routing.

Question 17

What is a security group?

Accepted Answer

A security group is a stateful virtual firewall attached to instances, allowing inbound and outbound traffic by protocol, port, and source. Being stateful, return traffic is automatically permitted. It is a primary control for least-privilege network access, complementing subnet-level network ACLs.

Question 18

What does IAM do in the cloud?

Accepted Answer

IAM manages authentication and authorization, defining users, groups, roles, and policies that grant least-privilege permissions to cloud resources. Roles let services and instances assume temporary credentials without long-lived secrets, making IAM central to cloud security and the principle of least privilege.

Question 19

What is serverless computing?

Accepted Answer

Serverless (functions as a service like Lambda) runs your code in response to events without provisioning or managing servers, auto-scaling to zero and billing per invocation and duration. It speeds development and cuts idle cost, with tradeoffs in cold starts, execution limits, and vendor coupling.

Question 20

What is a managed database service?

Accepted Answer

A managed database (RDS, Cloud SQL) offloads provisioning, patching, backups, replication, and failover to the provider, so teams focus on schema and queries instead of operations. It improves reliability and reduces toil, trading some configuration control and added cost versus self-managed databases.

Question 21

When would you choose NoSQL over a relational database?

Accepted Answer

Relational (SQL) databases enforce schemas and ACID transactions, ideal for structured, related data and complex queries. NoSQL stores (key-value, document, wide-column, graph) trade some consistency or joins for flexible schemas and horizontal scale, suiting high-volume, evolving, or denormalized workloads. Choice depends on access patterns.

Question 22

Why use a cache in cloud architecture?

Accepted Answer

Caching stores frequently accessed data in fast memory (Redis, Memcached, CDN) so repeated reads avoid slower databases or computation, cutting latency and backend load. Key concerns are invalidation, time-to-live, and consistency, since stale cache entries can serve outdated data.

Question 23

What role does DNS play in cloud applications?

Accepted Answer

DNS resolves human-readable domain names to IP addresses, directing users to the right endpoints. Cloud DNS (Route 53) adds routing policies like latency-based, geolocation, weighted, and failover routing, plus health checks, making it a key tool for global traffic management and high availability.

Question 24

What does an API gateway provide?

Accepted Answer

An API gateway is a managed front door for APIs that handles routing, authentication, rate limiting, request/response transformation, caching, and monitoring. It decouples clients from backend services (often microservices or serverless functions) and centralizes cross-cutting concerns.

Question 25

What is a message queue used for?

Accepted Answer

A message queue (SQS, RabbitMQ) buffers messages so producers and consumers operate independently and asynchronously, smoothing load spikes and improving resilience. If a consumer is down, messages wait; this decoupling enables reliable, scalable, loosely coupled architectures.

Question 26

How does publish-subscribe messaging work?

Accepted Answer

In pub/sub, publishers send messages to a topic without knowing subscribers, and the system delivers each message to all interested subscribers. It enables one-to-many, event-driven fan-out and loose coupling, contrasting with point-to-point queues where each message goes to a single consumer.

Question 27

What does high availability mean?

Accepted Answer

High availability means a system remains accessible with minimal downtime by eliminating single points of failure through redundancy across instances and availability zones, health checks, and automatic failover. It is measured in nines (99.9 percent, 99.99 percent) of uptime and drives multi-AZ designs.

Question 28

How does fault tolerance differ from high availability?

Accepted Answer

Fault tolerance means the system continues operating without interruption even when components fail, typically via full redundancy that masks failures instantly. High availability tolerates brief recovery time. Fault tolerance is stronger and costlier; many systems target high availability as a pragmatic balance.

Question 29

What is disaster recovery in the cloud?

Accepted Answer

Disaster recovery is the strategy and tooling to recover applications and data after a major failure or regional outage, defined by RTO (how fast to recover) and RPO (how much data loss is acceptable). Strategies range from backup-and-restore to pilot light, warm standby, and active-active multi-region.

Question 30

Why are regular backups important even in the cloud?

Accepted Answer

Backups create recoverable copies of data to protect against deletion, corruption, ransomware, or human error, which cloud durability alone does not prevent. Good practice follows retention policies, tests restores, and stores copies across regions or accounts to survive localized failures.

Question 31

What is a storage snapshot?

Accepted Answer

A snapshot captures the state of a volume at a moment in time, often incrementally storing only changed blocks for efficiency. Snapshots enable backups, fast volume cloning, and recovery to a known state, and can be copied across regions for disaster recovery.

Question 32

What is an SLA?

Accepted Answer

An SLA is a formal commitment to a measurable service level, most commonly uptime (e.g. 99.9 percent), with remedies like credits if missed. Architects combine the SLAs of dependent services to estimate overall availability and design redundancy to meet their own targets.

Question 33

What is latency?

Accepted Answer

Latency is the delay between sending a request and receiving a response, driven by network distance, processing time, and queuing. Low latency matters for interactive and real-time systems; it is reduced with CDNs, edge locations, regional deployment near users, and caching.

Question 34

What is throughput in a system?

Accepted Answer

Throughput is the rate of successful work, such as requests per second or bytes per second. It is distinct from latency: a system can have high throughput yet high latency. Scaling out, batching, and parallelism improve throughput, which is a key capacity-planning metric.

Question 35

How can you optimize cloud costs?

Accepted Answer

Cloud cost optimization includes right-sizing instances, using reserved or savings plans and spot instances, autoscaling, deleting idle resources, choosing appropriate storage tiers, and monitoring spend with budgets and tagging. The goal is matching capacity and pricing model to actual demand.

Question 36

What is the pay-as-you-go pricing model?

Accepted Answer

Pay-as-you-go bills for actual usage (compute hours, storage, requests) with no upfront commitment, converting capital expense to operating expense and enabling experimentation. It is flexible but can grow costly at steady scale, where reserved or committed-use discounts are cheaper.

Question 37

When do reserved instances save money?

Accepted Answer

Reserved instances (or savings plans) offer significant discounts versus on-demand in exchange for a one or three year commitment to a certain usage. They suit steady, predictable baseline workloads; bursty or short-lived needs are better served by on-demand or spot pricing.

Question 38

What are spot instances good for?

Accepted Answer

Spot instances sell spare capacity at steep discounts but can be reclaimed with little notice, so they suit fault-tolerant, stateless, or batch workloads (rendering, CI, big-data jobs) that can checkpoint and resume. They are unsuitable for stateful or always-on critical services.

Question 39

Why tag cloud resources?

Accepted Answer

Tags are key-value labels on resources used to organize them, allocate and report cost by team or project, automate operations, and enforce governance policies. A consistent tagging strategy is foundational for cost management, security, and lifecycle automation at scale.

Question 40

Why are monitoring and logging essential in the cloud?

Accepted Answer

Monitoring collects metrics (CPU, latency, errors) and triggers alerts, while logging records events for debugging and audit; together with tracing they form observability. They enable detecting incidents, diagnosing root causes, capacity planning, and meeting compliance in dynamic cloud environments.

Question 41

What is infrastructure as code?

Accepted Answer

Infrastructure as code (Terraform, CloudFormation) defines resources in declarative config that is version-controlled, reviewed, and applied automatically, making environments reproducible, auditable, and consistent. It replaces manual clicking, reduces drift, and enables rapid, reliable provisioning and disaster recovery.

Question 42

What do CI and CD mean?

Accepted Answer

Continuous integration automatically builds and tests every code change to catch issues early and keep the main branch healthy. Continuous delivery/deployment automates releasing those validated changes to environments, enabling frequent, reliable, low-risk deployments through an automated pipeline.

Question 43

Why are stateless services easier to scale?

Accepted Answer

Stateless services keep no client session data locally, so requests can go to any instance, making horizontal scaling, load balancing, and instance replacement trivial. Stateful services hold session or data, requiring sticky sessions or external state stores, which complicates scaling and failover.

Question 44

What is the difference between encryption at rest and in transit?

Accepted Answer

Encryption at rest protects stored data on disks or in databases so stolen media is unreadable, typically via provider-managed or customer keys. Encryption in transit (TLS) protects data moving across networks from interception. Both are standard security baselines, often required for compliance.

Question 45

What is the cloud shared responsibility model?

Accepted Answer

Under the shared responsibility model, the provider secures the underlying infrastructure (hardware, facilities, managed service internals) while the customer secures their data, identities, access configuration, OS patching (for IaaS), and application security. Misunderstanding this split is a common cause of breaches.

Question 46

What is multi-tenancy in cloud services?

Accepted Answer

Multi-tenancy means many customers share the same physical infrastructure while being logically isolated, which drives the cloud's cost efficiency and scale. Providers enforce isolation through virtualization and access controls; tenants rely on this boundary, making isolation a core security concern.

Question 47

What is edge computing?

Accepted Answer

Edge computing runs computation close to users or devices (at edge locations or on-device) rather than a distant central region, cutting latency and bandwidth and enabling real-time and offline-tolerant use cases like IoT and gaming. It complements central cloud for data aggregation and heavy processing.

Question 48

What does provisioning mean in cloud computing?

Accepted Answer

Provisioning is allocating and configuring cloud resources (compute, storage, network) to a ready state. In the cloud it is fast and programmable, increasingly done via infrastructure as code and automation rather than manual setup, enabling on-demand, repeatable environment creation.

Question 49

What characterizes a microservices architecture?

Accepted Answer

Microservices decompose an application into small, focused services that are independently developed, deployed, and scaled, communicating over APIs or messaging. This enables team autonomy, targeted scaling, and fault isolation, at the cost of added operational and networking complexity versus a monolith.

Question 50

What is failover?

Accepted Answer

Failover automatically redirects traffic or promotes a standby (another instance, AZ, or region) when the primary component fails, minimizing downtime. It relies on health checks and redundancy and is central to high availability and disaster recovery designs.

Question 51

What is the purpose of a Well-Architected Framework?

Accepted Answer

A Well-Architected Framework provides best-practice pillars (operational excellence, security, reliability, performance efficiency, cost optimization, and sustainability) to review and improve cloud architectures. It offers questions and tradeoff guidance so teams can identify risks and design for the qualities that matter to their workload.

Question 52

Why deploy across multiple regions?

Accepted Answer

Multi-region architecture replicates workloads across regions to survive a full region outage, serve users with low latency worldwide, and meet data-residency rules. It adds complexity in data replication, consistency, traffic routing, and cost, so it is reserved for high-availability or global needs.

Question 53

What is the difference between active-active and active-passive?

Accepted Answer

In active-active, multiple regions or instances serve live traffic simultaneously, sharing load and failing over instantly, but require multi-master data handling and conflict resolution. In active-passive, a standby stays idle (or warm) and takes over on failure, which is simpler and cheaper but has failover delay and idle cost.

Question 54

What do RTO and RPO define?

Accepted Answer

Recovery Time Objective (RTO) is the maximum tolerable time to restore service after an incident; Recovery Point Objective (RPO) is the maximum acceptable amount of data loss measured in time. Together they set how aggressive (and costly) a disaster recovery strategy must be, from backups to active-active.

Question 55

How does a blue-green deployment work?

Accepted Answer

Blue-green keeps two production environments: blue (current) and green (new release). You deploy and test on green, then switch traffic over, enabling instant rollback by switching back. It minimizes downtime and risk but doubles environment cost during the transition and needs care with shared state like databases.

Question 56

What is a canary deployment?

Accepted Answer

A canary deployment routes a small percentage of traffic to the new version while monitoring metrics, then gradually increases the share if healthy or rolls back on problems. It limits blast radius and surfaces issues with real traffic, trading slower rollout for safety.

Question 57

What is immutable infrastructure?

Accepted Answer

Immutable infrastructure means servers are never modified after deployment; changes are made by building a new image and replacing instances. This eliminates configuration drift, makes deployments predictable and easy to roll back, and pairs naturally with infrastructure as code and autoscaling.

Question 58

What does a NAT gateway do?

Accepted Answer

A NAT gateway allows instances in a private subnet to initiate outbound internet connections (for updates or APIs) while preventing inbound connections from the internet, preserving their isolation. It is managed and scalable, and traffic through it incurs data-processing and egress costs.

Question 59

When would you use a transit gateway over VPC peering?

Accepted Answer

VPC peering connects two VPCs directly and is simple for a few networks, but it is non-transitive, so a full mesh grows quadratically. A transit gateway acts as a central hub connecting many VPCs and on-premises networks with transitive routing, simplifying large topologies at additional cost.

Question 60

What problem does a service mesh solve?

Accepted Answer

A service mesh (Istio, Linkerd) adds a dedicated infrastructure layer, usually via sidecar proxies, that handles inter-service traffic management, mutual TLS, retries, load balancing, and observability without changing application code. It is valuable in large microservice systems but adds operational complexity and overhead.

Question 61

What is the circuit breaker pattern?

Accepted Answer

A circuit breaker monitors calls to a dependency and, after repeated failures, trips open to fail fast instead of waiting on timeouts, protecting the caller and giving the dependency time to recover. It then half-opens to test recovery before closing. It prevents cascading failures and resource exhaustion.

Question 62

Why is idempotency important in distributed systems?

Accepted Answer

An idempotent operation produces the same result whether executed once or many times, so clients can safely retry after timeouts or network failures without double-charging or duplicating records. It is often implemented with idempotency keys and is essential for reliable at-least-once messaging.

Question 63

What is eventual consistency?

Accepted Answer

Eventual consistency guarantees that, absent new updates, all replicas will converge to the same value, but reads may temporarily return stale data. It trades immediate consistency for availability and scale, suiting use cases like social feeds or DNS where brief staleness is acceptable.

Question 64

What does the CAP theorem state?

Accepted Answer

The CAP theorem says a distributed data store cannot simultaneously guarantee consistency, availability, and partition tolerance; since partitions are unavoidable, you trade consistency for availability (AP) or availability for consistency (CP) during partitions. It guides database and architecture choices under failure.

Ready to practice the full interview?