AWS Solutions Architect (SAA-C03) - Storage & Compute Flashcards

Question 1

Amazon EBS vs. Amazon EFS

Accepted Answer

• **Amazon EBS (Elastic Block Store):** Block-level storage attached to a single EC2 instance within the **same Availability Zone** (exception: Multi-Attach for `io1`/`io2`). High performance, low latency.
• **Amazon EFS (Elastic File System):** POSIX-compliant, file-level storage shared across **multiple Availability Zones** and hundreds of EC2 instances simultaneously via NFS.

Question 2

S3 Glacier Storage Classes & Lifecycle Rules

Accepted Answer

Automates moving objects between storage classes to optimize cost:
• **S3 Glacier Instant Retrieval:** Millisecond retrieval for rarely accessed data.
• **S3 Glacier Flexible Retrieval:** 1–5 min (Expedited), 3–5 hrs (Standard), 5–12 hrs (Bulk).
• **S3 Glacier Deep Archive:** 12 hrs (Standard), 48 hrs (Bulk); lowest cost storage.
• *Rule Requirement:* Objects must transition in sequence (e.g., Standard $\rightarrow$ Standard-IA $\rightarrow$ Glacier Flexible $\rightarrow$ Deep Archive).

Question 3

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

Accepted Answer

• **ALB (Layer 7):** HTTP/HTTPS traffic. Supports path-based routing, host-based routing, query string filtering, and containerized workloads (`ECS`). Slow compared to NLB.
• **NLB (Layer 4):** TCP/UDP/TLS traffic. Handles millions of requests per second with ultra-low latency. Assigns a **Static IP** per AZ and supports Elastic IPs.

Question 4

NAT Gateway vs. NAT Instance

Accepted Answer

• **NAT Gateway:** Fully managed by AWS, highly available within a single AZ (scales up to 100 Gbps automatically), requires no security groups or OS maintenance.
• **NAT Instance:** Self-managed EC2 instance in a public subnet; requires manual scaling, security group management, and software patching.

Question 5

AWS Transit Gateway

Accepted Answer

A regional network transit hub used to connect multiple **VPCs** and **on-premises networks** (via Direct Connect or VPN) using a hub-and-spoke topology. Eliminates the complex, transitive peering relationships required by traditional VPC peering.

Question 6

Recovery Time Objective (RTO) vs. Recovery Point Objective (RPO)

Accepted Answer

• **RTO (Recovery Time Objective):** Maximum acceptable **downtime** after a disaster before service is restored (time-based measure).
• **RPO (Recovery Point Objective):** Maximum acceptable amount of **data loss** measured in time (e.g., maximum age of files to recover from backup).

Question 7

Disaster Recovery (DR) Strategies (Cost vs. RTO/RPO)

Accepted Answer

1. **Backup & Restore:** Highest RTO/RPO; low cost (backups in S3).
2. **Pilot Light:** Critical core components run continuously (e.g., active DB replication); compute created on demand via templates.
3. **Warm Standby:** Scaled-down version of full environment running continuously.
4. **Multi-Site Active-Active:** Zero/near-zero RTO/RPO; highest cost (traffic served in real-time across multiple regions).

Question 8

EC2 Placement Groups

Accepted Answer

• **Cluster:** Packs instances close together inside a single AZ. Delivers low latency and high network throughput.
• **Spread:** Places instances on distinct hardware racks (max 7 instances per AZ). Used for critical applications that need hardware isolation.
• **Partition:** Spreads instances across logical partitions that do not share underlying hardware. Ideal for large distributed workloads like Hadoop or Kafka.

Question 9

Auto Scaling Group (ASG) Scaling Policies

Accepted Answer

• **Target Tracking:** Keeps a metric at a set target (e.g., maintain average CPU utilization at 50%).
• **Step Scaling:** Scales based on a set of threshold step adjustments (e.g., add 2 instances if CPU > 70%, add 4 if CPU > 85%).
• **Simple Scaling:** Scales based on a single metric threshold with a mandatory cooldown period.
• **Predictive Scaling:** Uses machine learning to forecast traffic patterns and scale proactively.

Question 10

S3 Bucket Policies vs. IAM Policies

Accepted Answer

• **S3 Bucket Policy:** Resource-based policy attached directly to an S3 bucket. Grants permissions across different AWS accounts or secures public access.
• **IAM Policy:** Identity-based policy attached to an IAM User, Group, or Role defining what AWS resources that identity can access.

Question 11

S3 Object Lock Options: Governance vs. Compliance Mode

Accepted Answer

Prevents object deletion or modification (WORM - Write Once, Read Many):
• **Governance Mode:** Users with specific IAM permissions (e.g., `s3:BypassGovernanceRetention`) can bypass or alter retention settings or delete the object.
• **Compliance Mode:** **No user**, including the AWS root user, can overwrite or delete the object or alter retention settings until the retention period expires.

Question 12

EBS Volume Types Quick Reference

Accepted Answer

• **General Purpose SSD (`gp2` / `gp3`):** Cost-effective, balances price and performance for transactional workloads.
• **Provisioned IOPS SSD (`io1` / `io2`):** Highest performance, low-latency storage for mission-critical databases.
• **Throughput Optimized HDD (`st1`):** Low cost for frequently accessed, throughput-intensive workloads (e.g., Big Data, log processing).
• **Cold HDD (`sc1`):** Lowest cost for infrequently accessed workloads with large data sets.

Question 13

VPC Public vs. Private Subnets

Accepted Answer

• **Public Subnet:** Route table contains an explicit route to an **Internet Gateway (IGW)** (`0.0.0.0/0` $\rightarrow$ `igw-xxxx`). Instances can have public IPs.
• **Private Subnet:** Route table does NOT route directly to an IGW. Accesses the internet out-bound only via a **NAT Gateway** located in a public subnet.

Question 14

Multi-Tier High Availability Architecture Pattern

Accepted Answer

• **Web Tier:** ALB forwarding traffic to an ASG of web instances across multiple AZs.
• **App Tier:** Internal ALB forwarding traffic to an ASG of application instances across multiple AZs.
• **DB Tier:** Multi-AZ RDS deployment (Primary instance with synchronous replication to a Standby in a secondary AZ for automatic failover).

Question 15

Amazon S3 Replication (SRR vs. CRR)

Accepted Answer

Replicates objects asynchronously across buckets:
• **SRR (Same-Region Replication):** Used for log aggregation, in-region data compliance.
• **CRR (Cross-Region Replication):** Used for latency reduction, cross-region disaster recovery.
• *Requirement:* **Versioning** must be enabled on both source and destination buckets.

Question 16

Amazon EFS Storage Classes & Lifecycle Management

Accepted Answer

• **EFS Standard:** Multi-AZ availability for active files.
• **EFS Infrequent Access (IA):** Lower storage cost for files not accessed daily; read/write fee applies upon access.
• **EFS One Zone:** Single AZ storage for non-critical or cost-sensitive data.
• **Lifecycle Management:** Automatically transitions files to IA after a designated period (e.g., 30 days without access).

Question 17

VPC Endpoints: Interface Endpoints vs. Gateway Endpoints

Accepted Answer

Enables private connections to supported AWS services without an IGW or NAT Gateway:
• **Gateway Endpoints:** Serves as a route target in route tables. Supports **Amazon S3** and **DynamoDB** only. Free of charge.
• **Interface Endpoints (AWS PrivateLink):** Uses an Elastic Network Interface (ENI) with a private IP address within your subnet. Supports most AWS services; charged per hour and data processed.

Question 18

AWS Launch Templates vs. Launch Configurations

Accepted Answer

• **Launch Templates:** Modern standard for configuring EC2 instances launched by Auto Scaling Groups. Supports **versioning**, parameter inheritance, T2/T3 Unlimited, and Spot Instance allocation strategies.
• **Launch Configurations:** Legacy feature; immutable (must create a new configuration to change parameters), lacks versioning.

Question 19

Gateway Load Balancer (GWLB)

Accepted Answer

Operates at Layer 3 (IP network layer). Listens on all IP packets and routes traffic to a fleet of inline third-party virtual network appliances (e.g., deep packet inspection firewalls, intrusion detection systems) before forwarding to the target destination.

Question 20

AWS Compute Optimizer

Accepted Answer

An AWS service that uses machine learning to analyze historical configuration and utilization metrics to recommend optimal AWS resources (EC2 instance types, EBS volumes, Auto Scaling groups, Lambda functions) to eliminate over-provisioning and cut costs.

AWS Solutions Architect (SAA-C03) - Storage & Compute

Cards in this deck

Related Flashcard Sets

MONASH MTH2010 - Multivariable Calculus & Vector Calculus

SYDNEY CHEM1111 - Chemical Bonding, Thermochemistry & Chemical Equilibrium

GCP Associate Cloud Engineer - Cloud IAM, Compute Engine & Networking

JLPT N4 - Intermediate Japanese Grammar, Particles & High-Frequency Kanji