AWS Transit GatewayVPC PeeringAWS PrivateLink
#1All three enable private connectivity between VPCs, so candidates pick based on which they associate with scale.
Deciding signal
VPC Peering creates a direct route between exactly two VPCs. Traffic does not traverse an intermediate hop, but peering is non-transitive — A-B and B-C does not create A-C. It has no bandwidth limit imposed by the service itself but requires a peering connection and route table entry for every pair. Transit Gateway is a regional hub: VPCs, VPNs, and Direct Connect attachments connect to it, enabling transitive routing between all attachments. It supports inter-Region peering and cross-account attachment. PrivateLink exposes a specific service endpoint (NLB-backed or VPCE service) to consumers in other VPCs or accounts — without granting full network routing access. When the scenario involves many VPCs needing transitive connectivity, Transit Gateway. When two specific VPCs need direct routing, Peering. When a service needs to be consumed privately without full VPC routing, PrivateLink.
Quick check
Is this direct routing between two VPCs (Peering), transitive hub-and-spoke across many VPCs (Transit Gateway), or exposing a specific service endpoint privately (PrivateLink)?
Why it looks right
Transit Gateway is the prominent ANS-C01-level answer for multi-VPC networking and candidates apply it to PrivateLink scenarios where only a service endpoint needs to be shared.
AWS Direct ConnectAWS Site-to-Site VPNAWS Client VPN
#2All three provide private access to AWS, so candidates default to Direct Connect as the "enterprise" answer.
Deciding signal
Site-to-Site VPN creates an encrypted IPsec tunnel between a customer gateway (on-premises router) and a Virtual Private Gateway or Transit Gateway — it connects entire networks over the public internet. Direct Connect provides a dedicated private network circuit from your data center to AWS, bypassing the public internet for consistent bandwidth and latency. Client VPN is OpenVPN-based and connects individual user devices (laptops, endpoints) to VPCs — it authenticates individual users, not networks. When the scenario describes an on-premises data center connecting to AWS with predictable latency and dedicated bandwidth, Direct Connect. When it describes employees or remote workers connecting to AWS resources from their devices, Client VPN.
Quick check
Is this a network-to-network connection with dedicated bandwidth (Direct Connect), an encrypted network-to-network tunnel over the internet (Site-to-Site VPN), or individual user devices connecting to a VPC (Client VPN)?
Why it looks right
Client VPN is frequently missed because candidates apply Site-to-Site VPN to all VPN questions. The signal is individual users or devices — not network-to-network.
Direct Connect GatewayTransit Gateway with Direct Connect attachment
#3Both extend Direct Connect connectivity, and the distinction requires understanding the attachment model.
Deciding signal
A Direct Connect Gateway allows a single Direct Connect connection (at a Direct Connect location) to reach VPCs across multiple AWS Regions and accounts. The VPCs must still be connected to the Direct Connect Gateway via Virtual Private Gateways — VPCs connected through it cannot route to each other through the Direct Connect Gateway (no transitive routing between VPCs). A Transit Gateway with a Direct Connect attachment (using a transit VIF) routes traffic between the on-premises network and all VPCs attached to the Transit Gateway, enabling transitive routing between VPCs in addition to on-premises access. When the scenario requires on-premises access to multiple Regions with no VPC-to-VPC routing through the Direct Connect path, Direct Connect Gateway. When transitive VPC-to-VPC routing through the on-premises path is needed, Transit Gateway with Direct Connect.
Quick check
Is the requirement to access VPCs in multiple Regions from a single Direct Connect without VPC-to-VPC routing (Direct Connect Gateway), or does the on-premises network need to be part of a transitive routing mesh (Transit Gateway with Direct Connect)?
Why it looks right
Direct Connect Gateway sounds like it handles all Direct Connect extension scenarios. Transit Gateway with Direct Connect is required when transitive routing between VPCs over the Direct Connect path is needed.
Security GroupsNetwork ACLsAWS Network Firewall
#4All three filter network traffic in VPCs, so candidates apply Security Groups as the default without checking scope.
Deciding signal
Security Groups are stateful and attach to ENIs — return traffic for allowed requests is automatically allowed without an explicit outbound rule. They evaluate all rules (no explicit deny; deny is implicit by omission). Network ACLs are stateless and apply at the subnet boundary — you must explicitly allow both inbound and outbound traffic, including return traffic. Rules are evaluated in order by rule number. AWS Network Firewall is a managed, stateful VPC firewall service deployed in a firewall subnet. It supports Suricata-compatible IPS/IDS rules, domain-based filtering, and protocol inspection — capabilities that Security Groups and NACLs cannot provide. When the scenario requires deep packet inspection, URL filtering, or IDS/IPS signatures across VPC traffic, Network Firewall.
Quick check
Is this per-instance stateful filtering (Security Groups), per-subnet stateless ACL rules (NACLs), or VPC-level deep packet inspection with IDS/IPS rules (Network Firewall)?
Why it looks right
Security Groups handle most traffic filtering scenarios and candidates apply them everywhere. Network Firewall is the right answer when IDS/IPS, domain filtering, or protocol-level inspection is explicitly required.
Route 53 Latency RoutingRoute 53 Geolocation RoutingRoute 53 Geoproximity RoutingRoute 53 Weighted Routing
#5All four Route 53 routing policies route traffic based on location or capacity, so candidates blur them in multi-region designs.
Deciding signal
Latency routing sends users to the AWS Region or endpoint with the lowest measured network latency — the decision is based on real-time latency, not geographic distance. Geolocation routing routes based on where the DNS query originates — country, continent, or default — regardless of latency. It is for legal or regulatory requirements (e.g., EU users must be served by EU endpoints). Geoproximity routing uses Traffic Flow and routes based on geographic distance, but you can apply a bias to expand or shrink the effective region for a specific endpoint. Weighted routing distributes traffic in proportions you define (e.g., 80/20 across two endpoints) regardless of geography or latency — used for canary deployments or A/B testing.
Quick check
Is this about serving users from the fastest endpoint (Latency), routing by user origin country for compliance (Geolocation), shifting traffic boundaries between regions (Geoproximity), or splitting traffic by percentage (Weighted)?
Why it looks right
Latency routing and Geolocation routing both involve geography and candidates conflate them. Geolocation is deterministic by user location; Latency is dynamic based on measured network performance.
NAT GatewayInternet GatewayEgress-Only Internet Gateway
#6All three enable communication between VPCs and the internet, so candidates apply Internet Gateway universally.
Deciding signal
An Internet Gateway enables bidirectional internet access for resources with public IPs in public subnets. NAT Gateway allows resources in private subnets to initiate outbound IPv4 connections to the internet without exposing those resources to inbound internet traffic — it is always deployed in a public subnet. An Egress-Only Internet Gateway serves the same purpose for IPv6 — it allows outbound-only IPv6 traffic from private resources but blocks inbound IPv6 connections initiated from the internet. When private resources need to download updates or reach internet services, NAT Gateway (IPv4) or Egress-Only IGW (IPv6). When public subnet resources need two-way internet access, Internet Gateway.
Quick check
Are private resources initiating outbound IPv4 connections (NAT Gateway), outbound IPv6 connections (Egress-Only IGW), or do public-subnet resources need bidirectional internet access (Internet Gateway)?
Why it looks right
Internet Gateway is the familiar service and candidates apply it to private subnet scenarios. NAT Gateway is required when resources must stay private but still reach the internet outbound.
All three are Elastic Load Balancing types and appear interchangeably in networking questions.
Deciding signal
ALB operates at Layer 7 and routes based on HTTP rules: path patterns, host headers, query strings, and HTTP methods. It terminates TLS and supports WebSocket. NLB operates at Layer 4 and handles TCP, UDP, and TLS at extreme throughput with static IP addresses and ultra-low latency. It preserves the client source IP address, which ALB does not by default. GWLB is not a general load balancer — it is purpose-built for inserting virtual network appliances inline. Traffic enters the GWLB, is forwarded to appliances (firewalls, IDS) via GENEVE tunneling, and returned to the GWLB for forwarding to the destination. When a scenario involves inline security appliances, GWLB is the specific answer.
Quick check
Is this Layer 7 content-based routing (ALB), high-throughput Layer 4 with static IPs (NLB), or transparent inline routing through security appliances (GWLB)?
Why it looks right
NLB and ALB are the commonly tested load balancers. GWLB is overlooked because its use case — inline appliance insertion — is a pattern candidates do not immediately associate with a load balancer.
Amazon Route 53 ResolverRoute 53 Hosted Zones
#8Both involve Route 53 and DNS, so candidates treat them as different names for the same functionality.
Deciding signal
Route 53 Hosted Zones are authoritative DNS zones — public (for internet DNS) or private (for VPC DNS). They store resource records (A, CNAME, MX, etc.) and respond to queries for your domains. Route 53 Resolver is the DNS resolver built into every VPC (runs at the VPC+2 address). Route 53 Resolver Endpoints (inbound and outbound) are used to integrate VPC DNS resolution with on-premises DNS servers: inbound endpoints let on-premises servers resolve Route 53 private hosted zone records; outbound endpoints let VPC resources forward DNS queries to on-premises DNS servers. When the scenario involves hybrid DNS resolution between on-premises and AWS VPCs, Resolver Endpoints. When it involves managing DNS records, Hosted Zones.
Quick check
Is this about storing and serving DNS records (Hosted Zones), or about forwarding DNS queries between on-premises and VPC environments (Resolver Endpoints)?
Why it looks right
Route 53 is associated with DNS record management, so candidates reach for Hosted Zones on all DNS questions. Resolver Endpoints are specifically for hybrid DNS forwarding and are frequently the correct ANS-C01 answer.
VPC Gateway EndpointVPC Interface EndpointAWS PrivateLink Service
#9All three provide private access to services, and "PrivateLink" is sometimes used loosely to describe all endpoint types.
Deciding signal
Gateway Endpoints are free, require only a route table entry, and work exclusively for S3 and DynamoDB. Interface Endpoints use PrivateLink to create ENIs with private IP addresses in your VPC and work for most AWS services (SSM, Secrets Manager, SQS, EC2 API, and many more). They have an hourly cost and support private DNS. PrivateLink Services (endpoint services) let you expose your own service (behind an NLB) to consumers in other VPCs or accounts — this is the publisher side of PrivateLink. When the scenario asks about consuming an AWS service privately, the answer is Gateway Endpoint (S3/DynamoDB) or Interface Endpoint (other AWS services). When a scenario describes sharing your own internal service across VPCs without VPC peering, PrivateLink Service is the answer.
Quick check
Is the target S3 or DynamoDB (Gateway Endpoint), another AWS service via ENI (Interface Endpoint), or exposing your own service to other VPC consumers (PrivateLink Service)?
Why it looks right
Interface Endpoints and PrivateLink Services are frequently confused because Interface Endpoints use PrivateLink technology. The distinction is direction: Interface Endpoints consume services; PrivateLink Services expose them.
AWS Network FirewallAWS WAFAWS Shield Advanced
#10All three protect AWS workloads from network threats and appear together in "defense in depth" questions.
Deciding signal
Network Firewall is deployed within VPCs and inspects all traffic passing through — not just HTTP. It supports stateful rules, IDS/IPS rule groups (Suricata-compatible), and domain-based filtering. It is for threats at the network and transport layers affecting all protocols. WAF is attached to CloudFront, ALB, API Gateway, or AppSync and inspects HTTP/HTTPS requests — it blocks web-layer threats like SQL injection, XSS, and rate abuse. Shield Advanced provides DDoS protection for elastic IPs, ELBs, CloudFront, Global Accelerator, and Route 53 — it absorbs volumetric attacks and provides access to the Shield Response Team. Each has a distinct threat model: VPC-level inspection (Network Firewall), HTTP application attacks (WAF), volumetric DDoS (Shield).
Quick check
Is this protecting against non-HTTP network threats within VPC traffic (Network Firewall), blocking malicious web requests at the application layer (WAF), or absorbing high-volume DDoS attacks (Shield Advanced)?
Why it looks right
WAF and Shield are the most visible security services for internet-facing workloads. Network Firewall is the answer when the scenario involves VPC-internal traffic, non-HTTP protocols, or IDS/IPS rule matching.
10 ANS-C01 questions. Pattern-tagged with trap analysis. Free, no signup required.
Start ANS-C01 Mini-Trainer →