AWS Certified Solutions Architect Associate (SAA-C03) - High Availability & Storage Flashcards

Question 1

S3 Standard-IA vs. S3 One Zone-IA

Accepted Answer

• **S3 Standard-IA**: High availability across $\ge 3$ Availability Zones for infrequently accessed data requiring millisecond retrieval.
• **S3 One Zone-IA**: Stores data in a single AZ, costing $20\%$ less than Standard-IA. Lower availability ($99.5\%$); data is destroyed if the AZ fails.

Question 2

S3 Glacier Storage Tiers: Instant vs. Flexible vs. Deep Archive

Accepted Answer

• **Glacier Instant Retrieval**: Millisecond access for data accessed $1-2$ times per quarter.
• **Glacier Flexible Retrieval**: Retrieval options from 1–5 minutes (Expedited), 3–5 hours (Standard), to 5–12 hours (Bulk).
• **Glacier Deep Archive**: Lowest cost storage tier; retrieval times range from 12 to 48 hours.

Question 3

S3 Intelligent-Tiering

Accepted Answer

Automatically optimizes storage costs by moving objects between frequent, infrequent, and archive access tiers based on usage patterns. Has no operational overhead or retrieval fees, making it ideal for unpredictable workloads.

Question 4

Amazon DynamoDB vs. Amazon Aurora (Architectural Choice)

Accepted Answer

• **Amazon DynamoDB**: Serverless NoSQL key-value database providing predictable single-digit millisecond latency at any scale. Best for unstructured data, simple queries, and high-throughput write spikes.
• **Amazon Aurora**: Fully managed relational database (MySQL/PostgreSQL compatible) with auto-scaling storage up to 128TB and low-latency read replicas. Best for complex SQL joins and ACID compliance.

Question 5

Amazon Aurora Read Replicas vs. Aurora Global Database

Accepted Answer

• **Read Replicas**: Up to 15 replicas within the *same AWS region* with sub-10ms replication lag.
• **Global Database**: Cross-region replication with typical latency under 1 second, providing disaster recovery with RTO $< 1$ minute and fast local reads worldwide.

Question 6

DynamoDB Accelerator (DAX) vs. DynamoDB Global Tables

Accepted Answer

• **DAX**: In-memory cache for DynamoDB delivering microsecond latency for read-heavy workloads without code rewrite.
• **Global Tables**: Fully managed multi-region, multi-active database offering cross-region replication for globally distributed applications.

Question 7

VPC Peering: Transitive Routing Constraint

Accepted Answer

VPC Peering connects two VPCs using AWS private network infrastructure. **Transitive routing is NOT supported**: if VPC A is peered with VPC B, and VPC B is peered with VPC C, VPC A *cannot* communicate with VPC C through VPC B.

Question 8

AWS Transit Gateway

Accepted Answer

A regional network transit hub that connects VPCs, on-premises networks (via VPN or Direct Connect), and AWS accounts in a central hub-and-spoke architecture. Unlike VPC Peering, Transit Gateway **supports transitive routing**.

Question 9

VPC Gateway Endpoints vs. Interface Endpoints (AWS PrivateLink)

Accepted Answer

• **Gateway Endpoints**: Free targets added directly to VPC route tables; supports **Amazon S3** and **DynamoDB** only.
• **Interface Endpoints**: Elastic Network Interfaces (ENIs) with private IP addresses powered by PrivateLink; supports most AWS services and third-party SaaS for hourly and per-GB charges.

Question 10

Route 53: Geolocation vs. Geoproximity Routing

Accepted Answer

• **Geolocation**: Routes traffic based on the *physical location of the user* (e.g., state, country, or continent).
• **Geoproximity**: Routes traffic based on the *geographic location of resources* and allows altering traffic flow boundaries using a positive/negative bias.

Question 11

Route 53: Latency vs. Weighted Routing

Accepted Answer

• **Latency Routing**: Directs users to the AWS region that delivers the lowest round-trip latency.
• **Weighted Routing**: Distributes incoming traffic across multiple resources based on assigned percentage weights (e.g., $80\%$ to endpoint A, $20\%$ to endpoint B for A/B testing).

Question 12

Route 53: Failover Routing Policy

Accepted Answer

Configures active-passive failover. Route 53 actively checks the health of the primary endpoint using health checks; if the primary becomes unhealthy, traffic automatically reroutes to the secondary (passive) disaster recovery target.

Question 13

RTO vs. RPO in Disaster Recovery

Accepted Answer

• **RTO (Recovery Time Objective)**: Maximum acceptable duration of system downtime after an outage before recovery.
• **RPO (Recovery Point Objective)**: Maximum acceptable amount of data loss measured in time (e.g., time elapsed since the last data backup).

Question 14

Disaster Recovery Strategy: Backup & Restore

Accepted Answer

• **Cost**: Lowest
• **RTO/RPO**: Highest (Hours to Days)
• **Architecture**: Data is backed up to S3/EBS snapshots. Infrastructure is fully created from infrastructure-as-code scripts (e.g., CloudFormation) only after a disaster occurs.

Question 15

Disaster Recovery Strategy: Pilot Light

Accepted Answer

• **Cost**: Low
• **RTO/RPO**: Medium (10s of Minutes to Hours)
• **Architecture**: Core backend data structures (e.g., databases) are continuously replicated in the recovery region. Application servers remain dormant as AMIs/launch templates and scale up only during a disaster.

Question 16

Disaster Recovery Strategy: Warm Standby

Accepted Answer

• **Cost**: Medium-High
• **RTO/RPO**: Low (Minutes)
• **Architecture**: A minimal, fully functional scaled-down version of the application runs continuously in the secondary region. Scales up to full production capacity during failover.

Question 17

Disaster Recovery Strategy: Multi-Site Active-Active

Accepted Answer

• **Cost**: Highest
• **RTO/RPO**: Near Zero
• **Architecture**: Production capacity runs simultaneously across multiple AWS regions. DNS (Route 53) routes global traffic dynamically to both active environments.

Question 18

Auto Scaling Policies: Target Tracking vs. Step Scaling

Accepted Answer

• **Target Tracking**: Automatically adjusts instance counts to maintain a specific target metric (e.g., keep average CPU usage at $50\%$).
• **Step Scaling**: Scales capacity in predefined steps based on metric thresholds (e.g., add $+2$ instances if CPU is $60-80\%$, add $+4$ instances if CPU $> 80\%$).

Question 19

Application Load Balancer (ALB) vs. Network Load Balancer (NLB)

Accepted Answer

• **ALB**: Operating at Layer 7 (HTTP/HTTPS), supports path-based/host-based routing, TLS termination, and containerized target groups.
• **NLB**: Operating at Layer 4 (TCP/UDP/TLS), handles millions of requests per second with ultra-low latency, and supports static/Elastic IPs per Availability Zone.

Question 20

Amazon EFS vs. EBS Multi-Attach

Accepted Answer

• **Amazon EFS**: Network-attached POSIX file system shared across thousands of EC2 instances simultaneously across multiple Availability Zones.
• **EBS Multi-Attach**: Attaches a single provisioned IOPS (`io1`/`io2`) EBS volume to multiple EC2 instances within the **same single Availability Zone** (requires cluster-aware file system).

Question 21

Amazon CloudFront vs. AWS Global Accelerator

Accepted Answer

• **CloudFront**: Content Delivery Network (CDN) optimizing performance by caching static and dynamic web content at global edge locations.
• **Global Accelerator**: Uses AWS global network and static Anycast IPs to route TCP/UDP traffic directly to regional endpoints, improving performance *without caching*.

AWS Certified Solutions Architect Associate (SAA-C03) - High Availability & Storage

Cards in this deck

Related Flashcard Sets

TypeScript - Advanced Type Systems, Generics & Design Patterns

Harvard ECON1011 - Intermediate Microeconomics: Game Theory & Market Structures

MIT 18.03 - Differential Equations: Ordinary Differential Equations, Laplace & Fourier

CPA Exam FAR - Financial Accounting & Reporting: US GAAP & Revenue Recognition