Study guide
5.3 Azure Firewall for Centralized Network Policy
Azure SC-900 Study Guide Exam
Track your progress as you read
Finished sections get a check mark.
1 Orientation: SC-900 Audience, Format, Scoring, Retakes, and Study Map
1.1 Credential Purpose and Audience 1.2 Exam Duration, Seat Duration, and Experience 1.3 Scoring, Scaled Score, and Pass-Rate Myths 1.4 Retake Policy and Certification Lifecycle 1.5 Scheduling, Price, and Delivery Options 1.6 Skills Measured and Study Map
2 Security, Compliance, and Identity Concepts: Zero Trust, Shared Responsibility, and GRC
2.1 Shared Responsibility Model 2.2 Defense in Depth and the CIA Triad 2.3 Zero Trust Model 2.4 Encryption, Hashing, and Data States 2.5 Governance, Risk, and Compliance 2.6 Identity Perimeter, Authentication, Authorization, and Federation
3 Microsoft Entra ID: Identities, Hybrid Identity, and Authentication
3.1 Microsoft Entra ID Foundations 3.2 Identity Types and Directory Objects 3.3 Hybrid Identity and Entra Connect Concepts 3.4 Authentication Methods and Sign-in Assurance 3.5 Multifactor Authentication 3.6 Password Protection, Self-Service Password Reset, and External Identities
4 Microsoft Entra Access Management, Governance, PIM, and Identity Protection
4.1 Conditional Access Purpose 4.2 Signals and Controls 4.3 Microsoft Entra Roles and Role-Based Access Control 4.4 Identity Governance, Entitlement Management, and Lifecycle Governance 4.5 Access Reviews and Privileged Identity Management 4.6 Microsoft Entra ID Protection and Risk
5 Azure Infrastructure Security: DDoS, Firewall, WAF, NSGs, Bastion, and Key Vault
5.1 Choose the Right Infrastructure Security Control 5.2 Azure DDoS Protection for Availability 5.3 Azure Firewall for Centralized Network Policy 5.4 Web Application Firewall for HTTP Workloads 5.5 Virtual Networks, Subnets, and Network Security Groups 5.6 Bastion and Key Vault for Administration and Secrets
6 Defender for Cloud: CSPM, Cloud Workload Protection, and Secure Score
6.1 Defender for Cloud Purpose and Exam Role 6.2 CSPM, Foundational Posture, and Defender CSPM 6.3 Policies, Standards, and Security Recommendations 6.4 Secure Score and Remediation Priority 6.5 Regulatory Compliance and Multicloud Posture 6.6 Cloud Workload Protection and Defender Plans
7 Microsoft Sentinel: SIEM, SOAR, Analytics, Hunting, Workbooks, and Playbooks
7.1 Sentinel as SIEM and SOAR 7.2 Data Connectors and Security Signals 7.3 Analytics Rules, Incidents, and Detection 7.4 Workbooks for Visual Review 7.5 Hunting, KQL Awareness, and Notebooks 7.6 Automation Rules and Logic Apps Playbooks
8 Microsoft Defender XDR: Office, Endpoint, Cloud Apps, Identity, Vulnerability, and Threat Intelligence
8.1 Defender XDR Suite and Microsoft Defender Portal 8.2 Defender for Office 365 8.3 Defender for Endpoint 8.4 Defender for Cloud Apps 8.5 Defender for Identity 8.6 Vulnerability Management and Threat Intelligence
9 Microsoft Purview, Service Trust, Privacy, Compliance Manager, and Compliance Score
9.1 Service Trust Portal and Assurance Resources 9.2 Privacy Principles and Microsoft Priva 9.3 Microsoft Purview Portal 9.4 Compliance Manager Assessments 9.5 Compliance Score and Improvement Actions 9.6 Scenario Selection for Trust, Privacy, and Compliance Manager
10 Microsoft Purview Data Classification, Labels, DLP, Retention, eDiscovery, and Audit
10.1 Data Classification Foundations 10.2 Content Explorer and Activity Explorer 10.3 Sensitivity Labels and Label Policies 10.4 Data Loss Prevention 10.5 Retention, Records Management, and Label Policies 10.6 eDiscovery, Audit, and Insider Risk Management
11 SC-900 Product Selection Playbooks and Common Microsoft Rebrand Traps
11.1 Read the Scenario Before Picking a Product 11.2 Entra Playbook for Identity, Access, and Governance 11.3 Defender Playbook for Protection, Detection, and Posture 11.4 Sentinel Playbook for SIEM, SOAR, and Hunting 11.5 Purview Playbook for Compliance, Data, and Risk 11.6 Renamed Products and Similar-Name Traps 11.7 Microsoft Learn Practice Assessment Review Loop
12 Final SC-900 Review, Timed Practice, Exam Day, and Next Certification Path
12.1 Domain-Weighted Final Review Plan 12.2 Timed Practice and Pacing 12.3 Exam Sandbox and Question Experience 12.4 Exam-Day Logistics and Boundaries 12.5 Scoring, Retakes, and After-Exam Decisions 12.6 Final Memory Check and Scenario Drills 12.7 Next Certification Path After SC-900
5.3 Azure Firewall for Centralized Network Policy
Key Takeaways Explain This Topic AI
Azure Firewall is a managed, cloud-native, fully stateful firewall-as-a-service with built-in high availability and autoscaling that inspects east-west and north-south traffic for an entire environment.
Threat intelligence-based filtering uses Microsoft's global feed to alert on or deny traffic to and from known malicious IPs and domains; FQDN tags and application rules allow or deny outbound access by domain name.
Three SKUs: Basic (small environments, L3-L4 + threat intelligence, up to ~250 Mbps), Standard (L3-L7, threat intel, DNS proxy, web categories, autoscale to ~30 Gbps), and Premium (adds TLS inspection, IDPS, and URL filtering).
Choose Azure Firewall over an NSG when the scenario needs centralized stateful policy, FQDN/application rules, threat intelligence, or a static outbound public IP; Firewall Manager manages Firewall, DDoS, and WAF policies at scale.
Last updated: June 2026
Centralize Firewall Enforcement in Azure
Azure Firewall is the SC-900 answer when a scenario asks for a managed network firewall service in Azure. Microsoft describes it as a cloud-native, intelligent, fully stateful firewall as a service for Azure cloud workloads, with built-in high availability and cloud scalability (autoscaling) . Being stateful means it tracks the state of each connection (the conversation), so a return packet for an allowed outbound flow is permitted automatically — it does not evaluate every packet in isolation. Those phrases separate it from simpler filtering tools.
Azure Firewall inspects both east-west and north-south traffic. East-west is traffic moving across workloads inside the environment — between subnets, virtual networks, or application tiers. North-south is traffic entering or leaving the environment — internet ingress or egress. When a question asks for central control over these paths, Azure Firewall is stronger than picking an NSG alone. It is typically deployed centrally in a hub virtual network (hub-and-spoke) and gives virtual-network resources a single static public IP for outbound traffic, which simplifies allow-listing on the destination side.
Threat Intelligence, FQDN Filtering, and SKUs
Two capabilities show up in exam wording. Threat intelligence-based filtering uses Microsoft's global threat feed to alert on or alert and deny traffic to and from known malicious IP addresses and domains. FQDN (fully qualified domain name) filtering — via application rules and FQDN tags — lets you allow or block outbound access by domain name (for example, allow *.windowsupdate.com ) instead of chasing dynamic IP addresses. Azure Firewall organizes its logic into network rules (IP/port/protocol), application rules (FQDN-based), and NAT rules (inbound DNAT).
Azure Firewall comes in three SKUs :
SKU Positioned for Notable capabilities
Basic Small/SMB or branch environments, lighter throughput (~250 Mbps) L3-L4 filtering + threat intelligence (alert mode)
Standard Most production workloads, autoscale to ~30 Gbps L3-L7 filtering, threat intelligence (alert/deny), DNS proxy, custom DNS, web categories
Premium Highly sensitive workloads (e.g., payment processing) Everything in Standard plus TLS inspection , IDPS (intrusion detection and prevention), and URL filtering
For SC-900 you do not memorize throughput numbers, but you should recognize that Premium adds TLS inspection and IDPS for deep, decrypted inspection, while Basic is the entry tier and Standard is the common middle. TLS inspection (Premium) terminates and decrypts outbound HTTPS, inspects it, then re-encrypts it — that is how Premium sees inside encrypted traffic; IDPS (Premium) detects and can block known attack signatures.
Firewall vs WAF vs NSG, and Firewall Manager
Do not confuse Azure Firewall with Web Application Firewall . Azure Firewall is broad network and application connectivity enforcement; it is not primarily a web-exploit inspection service. WAF is the service that protects web applications from threats such as SQL injection and cross-site scripting at layer 7.
Do not confuse Azure Firewall with a network security group either. NSGs filter traffic with simple 5-tuple rules bound to subnets or network interfaces. Azure Firewall is a managed firewall service usually placed centrally in a hub. In real architectures the two are used together : NSGs apply local subnet/NIC filtering, while Azure Firewall handles centralized inspection, FQDN/application rules, threat intelligence, and logging.
- Need in the prompt Why Azure Firewall fits
- Centralized application and network connectivity policy Firewall policy is created and enforced centrally
- Stateful firewall behavior Tracks connection state instead of per-packet evaluation
- Allow/deny outbound access by domain name FQDN filtering via application rules and FQDN tags
- Block traffic to known-malicious IPs/domains Threat intelligence-based filtering
- Consistent outbound source IP A static public IP for virtual-network resources
- Deep, decrypted inspection / signatures Premium SKU adds TLS inspection and IDPS
Azure Firewall Manager appears in broader scenarios: it is a security-management service for central security policy and route management across cloud perimeters. It can deploy and govern multiple Azure Firewall instances, associate DDoS protection plans , and manage WAF policies . The relationship to remember: Azure Firewall enforces policy; Firewall Manager manages policy at scale.
Choose Azure Firewall for central, managed, stateful firewall enforcement.
Choose WAF for web application exploit protection.
Choose NSGs for subnet or network-interface traffic filtering.
Use Firewall Manager wording as a scale-and-management clue.
Rule Types, DNAT, and a Worked Scenario
Azure Firewall processes traffic through three rule collections, and the order is exam-relevant because it is deterministic. DNAT rules are evaluated first (they translate and forward inbound internet traffic to internal resources). Network rules are next (filter by IP, port, and protocol — the layer 3/4 logic). Application rules are last (filter outbound HTTP/HTTPS and FQDN-based access). If a network rule allows a flow, application rules are not also processed for it; the engine applies the highest-priority match.
- Rule collection Filters on Typical use
- DNAT Inbound destination IP/port translation Publish an internal service to a specific external port
- Network Source/destination IP, port, protocol Allow/deny L3-L4 traffic between networks
- Application FQDN, FQDN tags, web categories Restrict outbound web access by domain
Work a representative scenario. A spoke VNet of VMs needs to reach only Windows Update and an approved SaaS API, must never reach arbitrary internet sites, and all egress must flow through a central inspection point with logging. 0/0` traffic to the firewall. An NSG cannot do FQDN-based filtering and is not centralized, so it is the wrong tool here. Add threat intelligence-based filtering in alert-and-deny mode and the firewall will also block any flow to a known-malicious IP or domain regardless of the allow rules.
This is the shape of many SC-900 firewall questions: the presence of FQDN/domain filtering , centralized egress , threat intelligence , or hub-and-spoke wording is what tips the answer from NSG to Azure Firewall.