AWS ANS-C01 cheat sheet with final-review decision rules for hybrid routing, DNS, Transit Gateway, PrivateLink, load balancing, observability, optimization, security, encryption, and exam traps.
Use this AWS Certified Advanced Networking - Specialty (ANS-C01) Cheat Sheet after you have studied the domains and before timed practice. The goal is not to memorize every service name. The goal is to classify the stem, find the failing network layer, and reject attractive answers that solve the wrong problem.
I verified these current AWS exam facts on May 24, 2026.
| Item | Value |
|---|---|
| Exam | AWS Certified Advanced Networking - Specialty |
| Exam code | ANS-C01 |
| Questions | 65 total |
| Scoring | 50 scored and 15 unscored; unscored items are not identified |
| Time | 170 minutes |
| Passing score | 700 on a 100-1000 scaled score |
| Cost | 300 USD |
| Candidate profile | 5+ years networking experience, including 2+ years cloud and hybrid networking |
With 65 questions in 170 minutes, the average budget is about 2.6 minutes per question. Long BGP, DNS, and multi-account routing stems can consume more than that, so mark the slow ones and return.
| Domain | Weight | Final-review priority |
|---|---|---|
| Network Design | 30% | Edge, DNS, load balancing, logging requirements, hybrid routing, multi-account and multi-Region connectivity. |
| Network Implementation | 26% | Direct Connect, VPN, Transit Gateway, PrivateLink, DNS forwarding, automation, testing, and route validation. |
| Network Management and Operation | 20% | BGP maintenance, route priority, traffic analysis, flow logs, optimization, quotas, subnet capacity, and cost. |
| Network Security, Compliance, and Governance | 24% | WAF, Shield, Network Firewall, GWLB, audit evidence, CloudTrail, encryption, ACM, ACM PCA, IPsec, MACsec, DNSSEC. |
ANS-C01 is a path-reasoning exam. If two answers are technically valid, the better answer usually preserves segmentation, scales across accounts, proves the path, and gives operations evidence.
ANS-C01 questions usually ask whether a path works, scales, fails over, and leaves enough evidence for operations. Keep this stack in mind before choosing a service:
If an answer changes a route, firewall, endpoint, or DNS record without proving the return path and evidence source, it is usually incomplete for this specialty exam.
Network Design covers edge services, DNS, load balancing, logging, hybrid routing, and multi-account or multi-Region connectivity. Design from requirements: protocol, source, destination, accounts, Regions, DNS, failover, visibility, and security boundary.
Network Implementation covers hybrid routing, multi-account/VPC connectivity, DNS, automation, and connectivity testing. Implement the route domain and validate it with Route Analyzer, Reachability Analyzer, logs, and metrics.
Network Management and Operation covers maintaining routing and connectivity, analyzing traffic, and optimizing performance, reliability, and cost. Troubleshoot from recent change, route table, propagation, BGP, DNS, flow evidence, and the actual bottleneck.
Network Security, Compliance, and Governance covers secure flows, security validation, audit, confidentiality, certificates, and DNSSEC. Security answers must include traffic steering, policy boundary, encryption or certificate choice, logging, audit, and alerting.
| Stem evidence | Strong answer includes | Reject answers that |
|---|---|---|
| Hybrid prefix is unreachable | BGP advertisement, accepted prefixes, route propagation, return route, tunnel/DX metrics | Start with packet mirroring before checking routing evidence |
| VPC-to-VPC path fails | Source/destination route tables, TGW association/propagation, SG/NACL, return path | Change one security group without proving route domain |
| Private service access required | PrivateLink or VPC endpoint, endpoint policy, private DNS, target health | Use broad peering/TGW when only one service should be exposed |
| DNS works in one place but not another | Hosted-zone association, Resolver rule, inbound/outbound endpoint, forwarding path, query logs | Fix network routes when the failure is name resolution |
| Central inspection required | Inspection VPC, TGW route tables, appliance mode/symmetric routing, firewall logs | Place a firewall that traffic can bypass |
| Compliance needs proof | Flow logs, CloudTrail, Firewall/Route 53/ELB/CloudFront logs, retention, encryption, alarms | Provide a control with no audit or alerting path |
Use this for long connectivity, outage, or design stems.
flowchart TD
S["Scenario"] --> C["Name the constraint: security, cost, latency, scale, failover, evidence"]
C --> P["Identify source, destination, account, VPC, subnet, Region"]
P --> R["Check route domain: VPC route table, TGW, DX, VPN, peering, PrivateLink"]
R --> G["Check guardrail: SG, NACL, firewall, endpoint policy, IAM, SCP"]
G --> D["Check DNS: hosted zone, Resolver, forwarding, health, DNSSEC"]
D --> E["Choose evidence: flow logs, CloudTrail, metrics, health checks, packet mirroring"]
E --> A["Pick the smallest answer that satisfies the stated constraint"]
If an answer changes a service but never proves the route, return path, DNS answer, control point, or evidence source, keep looking.
When two choices both sound network-valid, keep the one that satisfies the whole path:
| Need | Strong first fit | Reject when |
|---|---|---|
| Central transitive routing across many VPCs/accounts | Transit Gateway | The design only has two simple VPCs and no hub/routing-domain requirement. |
| Global network policy and segmentation | AWS Cloud WAN | The scenario only needs one Region or a simple TGW pattern. |
| Private producer-consumer service exposure | PrivateLink | The consumer needs broad bidirectional network routing, not one service. |
| Simple private VPC-to-VPC routing | VPC peering | The design needs transitive routing, central inspection, or overlapping CIDRs. |
| Predictable hybrid bandwidth | Direct Connect | The scenario only needs quick temporary encrypted backup. |
| Encrypted hybrid tunnel | Site-to-Site VPN/IPsec | The requirement is only link-level DX encryption or application TLS. |
| Link encryption on supported DX cross-connect | MACsec | The requirement is end-to-end application encryption across multiple hops. |
| HTTP edge acceleration and caching | CloudFront | The protocol is non-HTTP TCP/UDP acceleration with static IP requirements. |
| Static Anycast IPs and fast regional failover | Global Accelerator | The requirement is cacheable web content behavior. |
| Appliance insertion | Gateway Load Balancer | The requirement is native managed firewall policy rather than third-party appliance scale. |
| Managed network firewalling | AWS Network Firewall | The problem is HTTP request filtering at the application layer. |
| Application-layer web protection | AWS WAF | The traffic is non-HTTP east-west or outbound network flow inspection. |
| Stem clue | Eliminate first | Keep in play |
|---|---|---|
| Many VPCs need transitive routing | Full-mesh VPC peering | Transit Gateway route tables, Cloud WAN, segmentation. |
| VPC A peers with B, B peers with C, A cannot reach C | More peering routes through B | VPC peering is not transitive; use TGW or another hub design. |
| New on-prem prefix is unreachable | Packet capture first | BGP advertisement, accepted prefixes, propagation, route table, return route. |
| Same prefix over DX and VPN | Assume equal priority | Route priority, prefix specificity, BGP attributes, failover policy. |
| Overlapping CIDRs | More propagation | Renumber, NAT, PrivateLink, or isolated route domains. |
| Route table near limit | Add every specific route | Summarize, reduce propagation, split route domains, redesign. |
| Backup path is unexpectedly preferred | Security group only | BGP local preference communities, AS path, prefix length, tunnel/link health. |
Routing questions reward discipline. Source, destination, prefix, route source, route priority, return path, and propagation matter before packet-level analysis.
| Question clue | Check first | Exam-safe reasoning |
|---|---|---|
| More specific prefix exists | Longest prefix match | A /24 route beats a /16 route before other preference logic matters. |
| Same prefix over multiple paths | Route source and BGP attributes | Prefix specificity, local preference/community, AS path, and propagation decide preferred path. |
| Traffic leaves correctly but response fails | Return route and stateful control | Firewalls, TGW route tables, on-prem routing, and asymmetric inspection can break return flow. |
| Failover does not happen | Health, BGP withdrawal, route propagation, DNS TTL | A backup circuit is not useful unless routing actually moves traffic. |
| Inspection path works one way | Symmetric routing/appliance mode | Stateful appliances need both directions through the same inspection path. |
| Route scale or limits are stressed | Summarization and route-domain design | Adding specific routes can solve one outage while creating a scaling problem. |
| Requirement | Strong answer |
|---|---|
| Fast setup, encrypted connectivity, lower predictable throughput | Site-to-Site VPN. |
| Dedicated predictable private connectivity | Direct Connect with redundant locations/connections. |
| DX consistency plus IPsec confidentiality | VPN over Direct Connect. |
| Active/passive DX path | BGP attributes, local preference communities, prefix specificity, and health. |
| SD-WAN integration | Transit Gateway Connect or an approved vendor integration pattern. |
| Hybrid DNS | Route 53 Resolver inbound/outbound endpoints and forwarding rules. |
| Hybrid testing | Reachability Analyzer, Route Analyzer, flow logs, tunnel/DX metrics, and controlled failover. |
Do not confuse private connectivity with encrypted connectivity. Direct Connect is private, but AWS states that DX traffic is not encrypted by default.
| Requirement | Strong answer |
|---|---|
| Public authoritative DNS | Route 53 public hosted zone, correct record type, TTL, alias where appropriate. |
| Private VPC names | Private hosted zone associated with the right VPCs and accounts. |
| Hybrid resolution from on-prem to AWS | Resolver inbound endpoint and forwarding from on-prem DNS. |
| AWS VPC to on-prem names | Resolver outbound endpoint and forwarding rule. |
| Multi-account private DNS | Hosted zone association, RAM sharing where appropriate, Resolver rules, and governance. |
| DNS failover | Route 53 health checks and routing policy that matches the requirement. |
| DNS integrity | DNSSEC where supported and operationally maintained. |
| DNS visibility | Resolver query logs, Route 53 logs, CloudWatch, and audit retention. |
DNS answers are often wrong because they fix routing instead of name resolution, or they fix public DNS when the failure is a private hosted-zone or Resolver rule problem.
| Stem clue | Better answer | Why |
|---|---|---|
| On-prem clients cannot resolve private AWS names | Resolver inbound endpoint plus on-prem conditional forwarding | Private hosted zone alone is not visible to on-prem DNS. |
| AWS workloads cannot resolve on-prem names | Resolver outbound endpoint plus forwarding rule | VPC DNS does not know on-prem zones automatically. |
| Same private zone needed across accounts | PHZ association/RAM sharing and governance | Duplicating zones can cause drift and split-brain answers. |
| Public failover is too slow | Health checks, routing policy, TTL awareness | DNS failover still depends on resolver caching/TTL behavior. |
| Need DNS integrity | DNSSEC operational plan | DNSSEC signs DNS data; it does not encrypt application traffic. |
| Need DNS evidence | Resolver query logs and Route 53 logging | Flow logs show IP flows, not DNS answer correctness. |
| Stem says | Start with |
|---|---|
| Host/path routing, HTTP headers, web app | ALB. |
| TCP/UDP/TLS, static IP per AZ, high-performance transport | NLB. |
| Firewall or inspection appliance fleet | GWLB. |
| Cacheable global HTTP content | CloudFront. |
| Static Anycast IPs, TCP/UDP acceleration, fast endpoint failover | Global Accelerator. |
| DNS-level weighted, latency, failover, or geolocation behavior | Route 53 routing policies. |
| Public web attack patterns | WAF with CloudFront, ALB, API Gateway, or another supported entry point. |
| DDoS resilience | Shield-aligned public endpoint architecture plus WAF where application-layer filtering matters. |
Match the control to the OSI layer. WAF does not inspect arbitrary TCP flows. NLB does not give HTTP host/path routing. Route 53 does not provide instant traffic movement or static Anycast IPs.
| Pattern | Use when | Watch for |
|---|---|---|
| Transit Gateway | Many VPCs/accounts, route domains, inspection, hub routing. | Associations, propagation, blackhole routes, appliance mode, return path. |
| VPC sharing | Central network account owns subnets; app accounts launch resources. | Ownership boundary, quotas, security group and route governance. |
| AWS RAM | Share TGW, DNS, or network resources across accounts. | Acceptance, scope, principal, and organization integration. |
| PrivateLink | Many consumers need one provider service privately. | Endpoint policies, service permissions, private DNS, target health. |
| Gateway endpoint | Private S3/DynamoDB access from VPC route tables. | Bucket/IAM policy conditions and route-table association. |
| Interface endpoint | Private access to supported AWS services or endpoint services. | ENI security groups, private DNS, endpoint policy, subnet IP usage. |
PrivateLink is service exposure. Transit Gateway is network routing. VPC peering is direct private routing without transitive behavior.
| Need to prove | Use |
|---|---|
| Who changed a route table, security group, or firewall policy | CloudTrail plus current configuration state. |
| Whether a flow was accepted or rejected | VPC Flow Logs or firewall logs. |
| Packet-level inspection from selected ENIs | VPC Traffic Mirroring with scoped filters and target appliance. |
| HTTP request details at edge or load balancer | CloudFront, WAF, or ALB logs. |
| Transit Gateway traffic and topology | TGW Flow Logs, Network Manager, Route Analyzer. |
| Near-real-time operational alert | CloudWatch metrics, metric filters, alarms, EventBridge. |
| Long-term audit archive and queries | S3 log archive, Athena/SIEM, retention, encryption, least-privilege access. |
| Organization-wide security posture | Firewall Manager, Config, Trusted Advisor, Security Hub, delegated admin. |
CloudTrail proves API changes. Flow logs prove network flow behavior. Traffic Mirroring is packet-level and should be targeted, not the default answer for every investigation.
flowchart TD
S["Connectivity symptom"] --> C["Define source, destination, port, protocol, account, VPC, Region"]
C --> R["Check routes and propagation"]
R --> P["Check policy controls: SG, NACL, endpoint, firewall, IAM/KMS where relevant"]
P --> D["Check DNS answer and resolver path"]
D --> E["Collect evidence: Reachability Analyzer, Route Analyzer, Flow Logs, metrics"]
E --> F["Change the smallest failing layer and retest"]
| Tool | Best use | Not the first choice for |
|---|---|---|
| Reachability Analyzer | Validate a modeled path through VPC resources | Packet payload inspection |
| Route Analyzer / Network Manager | Inspect TGW routing/topology intent | Application-layer failures |
| VPC Flow Logs | Accepted/rejected flow metadata | DNS answer content or packet payload |
| Traffic Mirroring | Targeted packet-level analysis | Broad always-on logging for every subnet |
| CloudTrail | API changes and actor evidence | Data-plane packet loss |
| CloudWatch metrics/alarms | Operational trend and alerting | Proving who changed configuration |
| Bottleneck | Better first move |
|---|---|
| EC2 throughput | Instance network performance, ENA, placement, load distribution, target health. |
| HPC or tightly coupled ML cluster | EFA and placement-aware architecture. |
| Large packets fail intermittently | MTU and jumbo-frame support across every hop. |
| Global user latency | CloudFront for cacheable HTTP; Global Accelerator for static-IP TCP/UDP acceleration. |
| NAT gateway cost for S3/DynamoDB | Gateway endpoint where applicable. |
| Cross-AZ data charges | Review load balancer, NAT gateway, endpoint, and appliance placement by AZ. |
| Auto Scaling cannot add instances | Available subnet IPs in the selected AZ/subnet. |
| Route scale pressure | Summarization, propagation control, route-domain redesign. |
Optimization is a tradeoff question. Name the bottleneck before choosing the service.
| Requirement | Strong answer |
|---|---|
| SQL injection or HTTP application attack | AWS WAF at a supported HTTP entry point. |
| Volumetric edge DDoS concern | AWS Shield with resilient public endpoint architecture. |
| Outbound or east-west inspection | Network Firewall, proxy, or GWLB appliance path with routing symmetry. |
| Workload-level allow list | Security groups. |
| Subnet-level stateless guardrail | Network ACLs. |
| Private AWS service access control | Endpoint policy, IAM, resource policy, and endpoint design. |
| Central firewall/security group governance | Firewall Manager with AWS Organizations. |
| Security incident alerting | Logs/metrics to CloudWatch, EventBridge/SNS/ticketing, and runbook. |
A firewall that traffic bypasses is not a control. Central inspection answers must include route steering and return-path compatibility.
| Traffic path | Strong control pattern | Evidence to include |
|---|---|---|
| Internet inbound HTTP(S) | CloudFront/ALB/API Gateway plus WAF and Shield-aligned design | WAF logs, access logs, CloudWatch alarms |
| Internet outbound from workloads | Network Firewall/proxy/GWLB appliance path with route steering | Firewall logs, flow logs, route tables |
| East-west VPC traffic | SGs, NACLs where needed, Network Firewall/GWLB if inspection is required | Flow logs, firewall logs, Reachability Analyzer |
| Private AWS service access | Gateway/interface endpoint plus endpoint/resource/IAM policy | Endpoint policy, CloudTrail, flow logs |
| Hybrid encrypted tunnel | IPsec VPN or VPN over Direct Connect | Tunnel metrics, BGP routes, logs |
| Organization-wide firewall governance | Firewall Manager delegated admin with Organizations | Policy scope, compliance findings, account coverage |
| Requirement | Strong answer |
|---|---|
| Public HTTPS to AWS service | TLS with ACM on a supported endpoint such as CloudFront, ALB, API Gateway, or NLB where appropriate. |
| Internal service trust | ACM PCA, private certificates, mutual TLS where required. |
| Hybrid network-layer encryption | IPsec Site-to-Site VPN. |
| DX path plus IPsec confidentiality | VPN over Direct Connect. |
| DX link-layer encryption | MACsec on supported dedicated Direct Connect connections or LAGs. |
| S3 must reject unencrypted transport | Enforce HTTPS with bucket/IAM policy condition. |
| DNS integrity | DNSSEC where supported; TLS still protects the application payload. |
| End-to-end HTTPS through load balancer | Check whether target-side TLS is also configured. |
Private does not automatically mean encrypted. Segmentation, routing, and firewalling are not substitutes for TLS, IPsec, MACsec, or a certificate-management design.
For “choose two” or “choose three” network questions, the correct set often contains different parts of one working path.
| If the stem needs | Look for both |
|---|---|
| Hybrid failover | Connectivity option plus routing/failover policy or health evidence. |
| Private AWS service access | Endpoint type plus IAM/resource/endpoint policy. |
| Central inspection | Inspection service plus route-table steering/return-path design. |
| DNS across hybrid networks | Resolver endpoint plus forwarding/association rule. |
| Encryption compliance | Encryption mechanism plus certificate/key/endpoint lifecycle. |
| Audit proof | Log source plus delivery/alert/retention mechanism. |
Do not select two services that both solve the same half of the problem while leaving routing, policy, DNS, or evidence missing.
| Stem clue | Better instinct |
|---|---|
| “least operational effort” | Prefer managed AWS-native service and fewer moving parts. |
| “security/compliance” | State the trust boundary, policy, encryption, logging, and evidence. |
| “cost optimized” | Look for transfer path, NAT, endpoint, caching, cross-AZ, or route-scale waste. |
| “high availability” | Identify failure domain, health check, routing behavior, and recovery proof. |
| “troubleshoot” | Start with recent change, route, control, DNS, metrics, logs, then packet capture. |
| “many accounts” | Prefer Organizations, RAM, delegated admin, Firewall Manager, Config, central logs. |
| “overlap” | PrivateLink, NAT, renumbering, or isolated route domains before more propagation. |
| Cue | Remember |
|---|---|
| Peering | Direct, non-transitive, simple, no hub. |
| Transit Gateway | Hub routing, segmentation, route tables, propagation, inspection. |
| PrivateLink | One service privately, not broad network reachability. |
| Direct Connect | Predictable private path, not encrypted by default. |
| VPN | IPsec tunnel, encrypted, usually lower throughput than DX patterns. |
| MACsec | DX Layer 2 link encryption, supported dedicated paths, not end-to-end app encryption. |
| Resolver | Hybrid DNS endpoint and forwarding glue. |
| Flow Logs | Flow metadata and accept/reject evidence, not packet payload. |
| CloudTrail | Who changed AWS configuration. |
| WAF | HTTP request filtering. |
| Network Firewall/GWLB | Routed flow inspection. |
| ACM PCA | Private certificate authority and internal trust. |
Use IT Mastery for exact ANS-C01 practice routing, spaced review when available, and close-answer explanations after this final review.
ANS-C01 networking questions are path questions: source, destination, route, security, DNS, return path, health, evidence, and constraint. Answer in that order, and many plausible distractors become obviously incomplete.