Academy/VMware vDefend Security for VCF 5.x Administrator (6V0-21.25)
VCF

VMware vDefend Security for VCF 5.x Administrator (6V0-21.25)

VCF 9.0add-on exams (6V0)

VMware's rebranded NSX security platform. Distributed Firewall, Gateway Firewall, IDS/IPS, NDR, NTA, Malware Prevention, Zero-Trust security. Administrator exam (VMware vDefend Security for VCF 5.x Administrator), 75 questions, 90 minutes, 70 % passing. The official exam title names VCF 5.x; this page teaches the vDefend 9.x implementation in VCF 9.0 and notes 5.x differences where they matter.

6V0-21.25
VCP
75
Questions
90m
Duration
70%
Pass Score
47
Objectives

Exam Blueprint Weights

Section titles, groupings and weights below are VCDX Academy study groupings, NOT the official Broadcom blueprint structure; they are kept because the quiz objective ids index them. For this exam Broadcom's guide DOES publish 17 weighted topics (for example vDefend Firewall Architecture 11%, NTA & NDR 11%) — listed with their weights in the official-topics table below. Always cross-check the official exam guide. Official exam guide ↗
Section 1 — vDefend Architecture
~20%
Section 2 — Micro-Segmentation
~30%High Weight
Section 3 — Threat Prevention
~20%
Section 4 — Network Detection and Response
~15%
Section 5 — Operations
~15%

Official exam guide topics and weights

Published by Broadcom in the exam guide (Exam guide (Last Updated 05/12/2025) topic table, read 2026-09-24 and re-verified live 2026-10-10 (all 17 titles and weights match); 17 topics sum to 100 %.). These are the official topics; the bars above are the Academy study groupings.

#Official topicWeight
01Private Cloud Data Center Security5%
02VMware vDefend Firewall Architecture11%
03VMware vDefend Firewall Management11%
04Lateral Protection with vDefend Distributed Firewall7%
05Shared Services Platform (SSP)2%
06Planning Application Segmentation with vDefend Security Intelligence4%
07Context Aware Firewall and Identity Firewall5%
08Protecting Container Workloads with vDefend Firewall4%
09Gateway Firewall7%
10Security Automation5%
11Security Operations2%
12Role-Based Access Control4%
13Troubleshooting4%
14Advanced Threat Prevention2%
15IDPS (Intrusion Detection and Prevention System)8%
16Malware Prevention Detection8%
17NTA (Network Traffic Analysis) & NDR (Network Detection and Response)11%
Total100%

Version Evolution

vDefend (formerly NSX Security / NSX Advanced Threat Prevention) consolidates Broadcom's network security stack into unified platform. Evolution from NSX-T 3.x: (1) DFW rule scale increased 10x (now handles 10K+ rules per domain vs 1K in earlier versions). (2) Distributed IDS/IPS has existed since NSX-T 3.0; 3.2 added behaviour-based signatures, Gateway IDS/IPS, Malware Prevention and NDR on the NSX Application Platform.

(3) In vDefend 9.x the Security Services Platform (SSP 5.x) runs Security Intelligence, NDR and Malware Prevention on premises (Connected or Disconnected/air-gapped mode); campaigns are mapped to MITRE ATT&CK. (4) vDefend replaces point products: NSX Distributed Firewall (DFW), NSX Gateway Firewall (GWF), NSX Advanced Threat Prevention (ATP). Key operational patterns: (5) False positive tuning critical for IDS/IPS rollout (30-day baseline in monitor mode recommended). (6) DFW performance depends on rule scope and Applied-To optimization (avoid "All" broadcasts).

(7) Zero-trust micro-segmentation requires crawl-walk-run approach (baseline→enforce→continuous monitor).

Learning Outcomes

  • Understand VMware vDefend Security for VCF 5.x concepts and architecture
  • Design DFW policies with security groups and Applied-To scope so that rule count and per-vNIC evaluation stay small; explain why a group-based rule and an equivalent set of VM-to-VM rules differ vastly in scale and operability.
  • Use time-based DFW rules with logging for compliance windows and incident response (temporary troubleshooting access that closes automatically).
  • Size firewall logging: estimate log volume from rule hit rate (about 200 bytes per entry), plan syslog bandwidth and retention, and log only the rules that need evidence.
  • Run an IDS/IPS rollout in detect-only mode first, analyse repetitive false positives, tune or disable the offending signatures per profile, then move to prevent mode.
  • Distinguish vDefend Security Intelligence (flow visibility and rule recommendations for segmentation planning) from NTA/NDR (behavioural threat detection and campaign correlation), and apply each at the right phase of a zero-trust rollout.

vDefend Architecture Overview#

vDefend consolidates NSX security features (formerly distributed across multiple products) into a unified network security platform. At its core, vDefend = NSX-T security rebranded with enhanced detection capabilities.

vDefend = NSX Security Features + Enhanced Detection:
All vDefend components (DFW, Gateway FW, IDS/IPS, NDR, NTA, Malware Prevention) operate natively within NSX infrastructure. No additional appliances required for DFW or IDS/IPS (unlike legacy firewall plugins). The architecture is purely software-defined, with kernel-level enforcement on every ESXi hypervisor.

vDefend Component Stack:

  1. Distributed Firewall (DFW): Kernel-level L2-L7 stateful firewall on every ESXi host. Enforces micro-segmentation rules at the vNIC level. Inspects east-west traffic between VMs without requiring traffic hairpinning through an external appliance.
  1. Gateway Firewall: Stateful inspection on NSX Edge nodes (Tier-0 and Tier-1 gateways). Protects north-south traffic entering/leaving the VCF environment. Supports URL filtering and TLS inspection for advanced threat prevention.
  1. IDS/IPS: Signature-based intrusion detection and prevention. Available in Distributed mode (per-host, east-west) and Gateway mode (per-edge, north-south). Uses Broadcom Threat Intelligence Cloud for signature updates.
  1. Malware Prevention: Multi-technique malware detection using ML, static analysis, and cloud-based dynamic analysis (sandboxing). Requires Guest Introspection (thin agent in the VM) plus a Malware Prevention service VM (SVM) on each host of an activated cluster (VCF 9.0.1 docs p.2728).
  1. Network Traffic Analysis (NTA): ML-based anomaly detection on network traffic (beaconing, data exfiltration, lateral movement) that feeds NDR campaigns. Application-flow discovery and rule recommendations for micro-segmentation planning are vDefend Security Intelligence (exam topic 06), not NTA.
  1. Network Detection & Response (NDR): ML-based behavioural threat detection running on the Security Services Platform (SSP; NSX Application Platform in older releases). Correlates IDS/IPS, malware and NTA anomaly events into attack campaigns mapped to MITRE ATT&CK. Optional cloud services (sandboxing, threat intelligence, data sharing) in Connected mode; Disconnected mode for air-gapped deployments.

Key Architectural Principles:

  • Hypervisor-native: All enforcement happens in the ESXi kernel, not in the guest OS. Cannot be bypassed by compromised VMs.
  • Distributed: Every host is a firewall. No chokepoint, no traffic steering, no bottleneck.
  • Policy follows workload: When a VM migrates (vMotion), its security policy moves with it. No reconfiguration required.
  • Closed-loop security: DFW prevents → IDS/IPS detects → NTA analyzes → NDR correlates → DFW responds. Integrated feedback cycle.

VCF Integration: vDefend is managed via NSX Manager (which is part of every VCF instance). DFW rules, IDS/IPS profiles, and NTA data are all accessible from the NSX Manager UI. VCF Operations integrates vDefend alerts into the unified operations dashboard for cross-component correlation (security + performance + availability in one view).

Evolution from NSX Security:

  • NSX-T 3.0/3.1: DFW, Gateway Firewall and Distributed IDS/IPS (host-based, east-west) already shipped.
  • NSX-T 3.2 (release notes): Distributed IDS/IPS gains behaviour-based signatures; Gateway IDS/IPS (tech preview), Gateway URL filtering and TLS inspection; Malware Prevention for the Distributed Firewall (E-W) and Gateway (sandboxing); NSX Intelligence, NSX Metrics, NSX Malware Prevention and NSX Network Detection and Response hosted on the on-prem NSX Application Platform (NAPP, Kubernetes). In 3.2 the NDR correlation itself ran in a Broadcom/VMware cloud region (US/EU).
  • NSX 4.x / VCF 5.x: the security features are licensed and branded as VMware vDefend (vDefend Firewall; vDefend Firewall with Advanced Threat Prevention); the 6V0-21.25 exam is titled vDefend Security for VCF 5.x.
  • vDefend 9.x / VCF 9.0: Security Services Platform (SSP 5.x) replaces NAPP as the platform that runs Security Intelligence, NDR and Malware Prevention on premises, with Connected mode (cloud sharing, sandboxing, threat intelligence) or Disconnected mode for air-gapped estates; vDefend DFW 1-2-3-4 four-stage prescriptive segmentation and, separately, Firewall Rule Analysis for rule-base hygiene.

Key Takeaways

  • vDefend = DFW + Gateway FW + IDS/IPS + Malware Prevention + NTA + NDR. All components run natively within NSX infrastructure — no additional appliances for DFW/IDS.
  • Hypervisor-native enforcement: DFW rules enforced at vNIC level in ESXi kernel. Cannot be bypassed by guest OS. Policy follows VM on vMotion.
  • NDR, Security Intelligence and Malware Prevention run on the on-prem Security Services Platform (SSP 5.x; NAPP before it). NDR can be activated in Disconnected mode for air-gapped estates; Connected mode adds cloud sandboxing, threat intelligence and data sharing. IDS/IPS and DFW enforce on the hosts and Edges without SSP.

Distributed Firewall (DFW) Deep Dive#

DFW is the foundation of vDefend. Every ESXi host runs a kernel module (VSIP — vSwitch InterPosition filter) that enforces firewall rules at the vNIC level, inspecting every packet entering or leaving a VM before it hits the virtual switch.

DFW Architecture — Kernel-Level Enforcement

The DFW operates at Layer 2-7, positioned between the VM's vNIC and the vSwitch port. This position is critical: traffic is inspected BEFORE it reaches the network, making it impossible for a compromised VM to bypass the firewall. Unlike perimeter firewalls that only inspect north-south traffic, DFW inspects east-west (VM-to-VM) traffic on the same host without hairpinning to an external appliance.

Data Plane Components:

  • VSIP Filter: Kernel module on each ESXi host. Processes rules per-vNIC. Stateful — tracks connection state (SYN/SYN-ACK/ACK) and allows return traffic automatically.
  • Connection Table: Per-host state table tracking active connections. Its capacity is finite (check the vDefend configuration maximums for the current value). When full, new connections are dropped — monitor flow counts with nsxcli on the host, in NSX Manager or in VCF Operations.
  • Rule Table: Per-vNIC rule set compiled from NSX Manager policies. Rules evaluated top-to-bottom within each category. First match wins.

Control Plane:

  • NSX Manager pushes compiled rule sets to each transport node (ESXi host).
  • Rules are organized into categories evaluated left to right: Ethernet (L2) → Emergency → Infrastructure → Environment → Application (vDefend Firewall 9.1, Distributed Firewall Categories). The default rule sits at the bottom of the Application category and allows all traffic out of the box.
  • Within each category, rules are evaluated by Section (ordered), then by Rule within each section.

DFW Rule Design Patterns

Applied-To Scope: Controls which vNICs process a rule. CRITICAL for performance:

  • Applied-To = "DFW" (all vNICs): Every vNIC evaluates the rule. Expensive at scale.
  • Applied-To = Security Group: Only vNICs in the group evaluate the rule. Recommended.
  • Best Practice: Always scope Applied-To to the narrowest group possible.

Rule Categories — When to Use Each:

  • Emergency: Block active threats immediately (highest priority). Example: quarantine compromised VM by IP.
  • Infrastructure: Protect management plane (vCenter, NSX Manager, ESXi). Example: allow only admin subnet to SSH to ESXi.
  • Environment: Enforce zone boundaries (Dev/Staging/Prod). Example: deny all traffic from Dev to Prod segments.
  • Application: Micro-segmentation rules per application tier. Example: allow Web→App on TCP 8080, App→DB on TCP 3306.
  • Default: Catch-all rule at the bottom. Typically set to "deny all" for zero-trust, or "allow all" during initial rollout.

Security Groups & Dynamic Membership:

  • Static Groups: Manually assigned VMs. Simple but doesn't scale.
  • Dynamic Groups: Membership based on criteria — VM name pattern, OS type, tags, segment membership. Auto-updates as VMs are created/moved/deleted.
  • Tag-Based Groups: Most flexible. Apply NSX tags to VMs via automation (VCF Automation, PowerCLI, API). Tags can represent: application tier (web/app/db), environment (dev/prod), compliance zone (PCI/HIPAA), business unit.

DFW Rule Optimization

Firewall Rule Analysis (Security Services Platform): analyses the existing DFW rule base on demand or on a schedule and reports seven anomaly types — Duplication, Contradiction, Redundancy, Shadowing, Consolidation, Permissiveness (an ANY-source, ANY-destination, ANY-service allow rule) and Irrelevance (a rule referencing an empty group or an IP-set-based Applied-To) — so rules can be fixed, consolidated or removed (VMware Validated Solutions 9.x p.251). It has no unused-rule category; find rules with no hits from the DFW rule hit counts.

vDefend DFW 1-2-3-4 is a different SSP feature: a four-stage prescriptive segmentation journey that follows lateral traffic patterns — Stage 1 Security Segmentation Assessment (segmentation score), Stage 2 Secure Infrastructure, Stage 3 Secure Environments, Stage 4 Secure Applications (VCF 9.1 docs, vDefend DFW Stage Prescriptive Segmentation Design).

Manual hygiene checks for any rule review: duplicate rules, rules shadowed by a broader rule above them, contradictory actions for the same traffic, "Any" in source/destination/service, rules that could be merged with groups, and rules with zero hits over the review period.

DFW Performance Considerations:

  • Rule Count Impact: Linear degradation with rule count per vNIC. Keep below 1,000 rules per vNIC for optimal performance.
  • Connection Table: Monitor utilization. At 80%+ capacity, new connections may be delayed or dropped.
  • Group Size: Dynamic groups with 10,000+ members cause membership computation overhead. Split into sub-groups.
  • Logging: Enable selectively. Full logging on all rules generates massive syslog volume (200 bytes/rule/hit).

Traceflow for DFW Validation:

NSX Traceflow injects a tagged packet between two endpoints and reports: which DFW rules it matched, whether it was allowed/dropped, the path through the network. Essential for validating micro-segmentation rules before enforcement.

Exclusion List: VMs that should bypass DFW entirely (e.g., NSX Manager, backup appliances). Add via NSX Manager → Security → Distributed Firewall → Exclusion List. Use sparingly — excluded VMs have no DFW protection.

Key Takeaways

  • Exam Focus: rule count matters. "Allow VM1 to VM2" and "allow prod-web group to prod-db group" can be semantically identical but differ vastly in scale and operability. Design policies with security groups (tags, dynamic membership) and a tight Applied-To, not per-VM rules — that is the difference between a policy that survives growth and one that collapses under its own rule count.
  • Exam Focus: Time-based rules are useful for compliance (enforce access restrictions) and incident response (temporarily open firewall for troubleshooting, auto-close after 1 hour). Combined with logging, they provide audit trail of network access patterns.
  • Log Storage Planning: Each firewall rule log entry ~200 bytes. At 10,000 packets/sec, that's 2MB/sec logs, or 172GB/day. For central syslog, ensure network bandwidth available (100Mbps recommended for large deployments). Use syslog rotation, compression, and archival to manage storage.

Gateway Firewall & North-South Security#

Gateway Firewall inspects traffic crossing Tier-0 and Tier-1 gateways. Provides N-S (north-south, ingress/egress) protection with stateful inspection and advanced threat features.

Gateway Firewall Architecture

Deployment: Gateway Firewall runs on NSX Edge nodes as part of the Tier-0 and Tier-1 gateway service routers. Unlike DFW (which is distributed across every ESXi host), Gateway FW is centralized on Edge nodes — all north-south traffic must transit through Edge nodes.

Traffic Flow: External network → Physical router → Tier-0 Gateway (Gateway FW inspects here) → Tier-1 Gateway (optional second inspection) → Overlay segment → VM.

Gateway FW vs DFW — When to Use Each:

  • Gateway FW: North-south perimeter enforcement. Inspects traffic entering/leaving the VCF environment. Use for: external-facing rules (block IPs, geo-filtering, protocol restrictions), NAT-aware inspection, TLS decryption for inbound services.
  • DFW: East-west micro-segmentation. Inspects traffic between VMs within the VCF environment. Use for: zero-trust policies, application tier isolation, lateral movement prevention.
  • Best Practice: Use BOTH. Gateway FW as perimeter defense. DFW as internal defense. Defense in depth.

Gateway Firewall Rule Categories:

The Gateway Firewall has its own category set, evaluated left to right: Emergency (quarantine/allow) → System (read-only, auto-generated for control-plane traffic such as VPN) → Pre Rules (rules shared by several gateways) → Gateway Specific (rules for one gateway) → Default (vDefend Firewall 9.1, Gateway Firewall Categories). Rules are applied per gateway (Tier-0 or Tier-1); Applied-To is the gateway or its interfaces.

Gateway IDS/IPS:

  • IDS/IPS can be enabled on Gateway Firewall rules (in addition to DFW rules).
  • Gateway IDS inspects north-south traffic at the edge — catches threats at the perimeter.
  • Performance consideration: IDS/IPS on Edge nodes reduces maximum throughput. Plan Edge sizing accordingly (Large Edge VM for IDS-enabled deployments).

Gateway Firewall Services:

  • Stateful Inspection: Connection tracking for TCP/UDP/ICMP with session timeout management.
  • URL Filtering: Category-based web filtering on Tier-1 gateways. Block access to known-malicious URLs, phishing sites, command-and-control domains.
  • TLS Inspection: Decrypt and inspect HTTPS traffic at the gateway. Requires CA certificate deployment to client VMs for re-encryption. Use selectively — impacts performance and requires certificate management.
  • FQDN-Based Rules: Allow/deny based on fully qualified domain name (e.g., allow *.broadcom.com, deny *.malware-domain.com), using DNS snooping to map FQDN to IP. Available on the Gateway Firewall and on the DFW (L7 DNS context-profile rule first, then the FQDN allow/deny rule) — not gateway-exclusive.

Gateway Firewall Scaling:

  • Each Edge node has a maximum connection table size and throughput limit.
  • For high-traffic environments: deploy Edge nodes in active-active (ECMP) mode behind the Tier-0 gateway.
  • IDS/IPS reduces throughput by 20-40% depending on signature count and traffic profile.
  • Monitor Edge CPU and memory via VCF Operations to detect saturation.

Exam Considerations:

  • Understand the difference between Gateway FW (centralized, north-south) and DFW (distributed, east-west).
  • Know that Gateway IDS/IPS is supported on BOTH Tier-0 and Tier-1 gateways.
  • URL filtering and TLS inspection are Gateway FW features — NOT available on DFW.
  • Gateway FW rules reference the specific gateway (Tier-0 or Tier-1) as the Applied-To scope.

Key Takeaways

  • Gateway FW = north-south perimeter defense on Edge nodes. DFW = east-west micro-segmentation on ESXi hosts. Use both for defense-in-depth.
  • Gateway FW-only features: URL filtering (L7 access profile on a Tier-1) and TLS inspection. FQDN filtering works on both the DFW (L7 DNS context profile + FQDN rule, DNS snooping) and the Gateway Firewall (vDefend Firewall 9.1, FQDN Filtering).
  • Gateway IDS/IPS supported on both Tier-0 and Tier-1 gateways. Reduces Edge throughput by 20-40% — factor into Edge sizing.

IDS/IPS & Signature-Based Detection#

IDS/IPS detects and prevents attacks using threat signatures distributed by Broadcom's Threat Intelligence Cloud. vDefend supports IDS/IPS on both Distributed Firewall (east-west, per-host) and Gateway Firewall (north-south, on Edge nodes).

IDS vs IPS — Operational Modes

IDS (Intrusion Detection System): Monitor-only mode. Detects malicious traffic and generates alerts but does NOT block. Traffic continues to flow. Use during initial deployment and tuning phase.

IPS (Intrusion Prevention System): Inline enforcement mode. Detects AND blocks malicious traffic. Dropped packets logged with signature details. Use after tuning in IDS mode to avoid false-positive-induced outages.

Deployment Architecture

Distributed IDS/IPS: Runs on each ESXi host as part of the VSIP kernel module. Inspects east-west traffic between VMs. Activated per-rule in DFW policy (add IDS/IPS profile to DFW rules). Each host processes signatures locally — no traffic steering to external appliance.

Gateway IDS/IPS: Runs on NSX Edge nodes (Tier-0 and Tier-1 gateways). Inspects north-south traffic entering/leaving the VCF environment. Activated per-rule in Gateway Firewall policy. Inspects traffic at the gateway before routing to external networks.

Key Difference: Distributed IDS catches lateral movement (attacker moving between VMs inside the datacenter). Gateway IDS catches perimeter threats (external attacks entering via internet/WAN). Both are needed for defense-in-depth.

Signature Management

Signature Sources: Broadcom Threat Intelligence Cloud delivers signature updates. NSX Manager downloads signatures on a configurable schedule (default: daily check). Signatures classified by:

  • IDS Severity: Critical, High, Medium, Low, Suspicious.
  • CVSS Score: 0-10 severity rating. Critical (9.0-10.0), High (7.0-8.9), Medium (4.0-6.9), Low (0.1-3.9).
  • Attack Type: Exploit, malware, C2 (command-and-control), recon, DoS.
  • Protocol: HTTP, DNS, SMB, SSH, TLS, custom protocols.

IDS/IPS Profiles: vDefend has no built-in Critical/Moderate/Aggressive profiles. An IDS/IPS profile is an administrator-defined grouping of signatures selected by filters — IDS severity (Critical, High, Medium, Low, Suspicious), CVSS score, attack type/target and product affected — with per-signature action overrides and exclusions (vDefend ATP 9.1 docs, Manage Signatures). Profiles are attached to Distributed or Gateway Firewall IDS/IPS rules. Design pattern: a "Critical + High" severity profile for production tiers, broader profiles (adding Medium/Low) for high-risk zones once tuned.

IDS/IPS Tuning Methodology

Phase 1 — Baseline (Days 1-30): Deploy in IDS (detect-only) mode across all segments. Monitor alerts via VCF Operations or syslog/SIEM. Identify false positives — legitimate traffic flagged as malicious.

Phase 2 — Tune (Days 30-45): Disable signatures generating false positives (whitelist by source/destination/signature ID). Adjust profiles per application tier. Create exclusions for known-good traffic patterns (e.g., vulnerability scanners, backup tools).

Phase 3 — Enforce (Day 45+): Switch from IDS to IPS mode tier-by-tier. Start with least-critical tier (dev/test). Monitor for blocked legitimate traffic. Gradually promote to production tiers.

Phase 4 — Continuous: Review new signature updates monthly. Re-tune after infrastructure changes (new applications, network redesign). Track metrics: detection rate, false positive rate, time-to-response.

Custom Signatures: Create custom IDS/IPS rules using Suricata rule syntax for organization-specific threats or zero-day patterns. Upload via NSX Manager → Security → IDS/IPS → Signatures → Custom.

Performance Impact: Distributed IDS/IPS adds CPU overhead per host (varies by signature count and traffic volume). Monitor ESXi host CPU via VCF Operations. If CPU exceeds 80% sustained, reduce signature count or scope IDS to specific segments only.

Key Takeaways

  • False Positive Management: the most common cause of IDS/IPS rollout failure. Strategy: (1) run in detect-only mode for about 30 days, (2) analyse repeated false positives, (3) tune or disable the offending signatures per profile, (4) switch the profile to prevent mode. Typical false positives: legitimate DNS tunnelling-like traffic, legacy encrypted protocols, monitoring and backup agents, vulnerability scanners.

Network Detection & Response (NDR)#

NDR is ML-powered behavioral threat detection. Unlike IDS/IPS (signature-based), NDR detects novel attacks by analyzing network behavior patterns, lateral movement, and command-and-control communications. NDR fills the gap where signature-based detection fails — zero-day exploits, living-off-the-land attacks, and encrypted threat channels.

NDR Architecture

NDR runs on the Security Services Platform (SSP), an on-prem Kubernetes-based platform deployed by the SSP Installer and onboarded to NSX Manager (NSX Application Platform/NAPP in NSX 3.2-4.x deployments):

Data Collection Layer:

  • Network sensors on each ESXi host collect metadata (NOT full packet capture): source/destination IP, port, protocol, byte count, packet count, timing, DNS queries, TLS certificate details.
  • Metadata is sent to the SSP, where aggregation, correlation and campaign building run. In Connected mode selected artefacts (file samples for sandboxing, telemetry for threat intelligence) go to the Broadcom cloud; in Disconnected mode nothing leaves the data centre.
  • No raw payload is stored or exported — metadata and flow summaries only.

Analytics Engine (on the SSP):

  • Aggregation Engine: Collects signals from IDS/IPS, NTA, and Malware Prevention. Combines individual alerts into correlated campaigns.
  • Correlation Engine: Links related events across time, hosts, and users. Example: DNS anomaly on VM-A at T1 → unusual outbound connection from VM-A at T2 → lateral movement to VM-B at T3 → data exfiltration attempt at T4. Individual events are low-confidence; correlated campaign is high-confidence.
  • Context Engine: Enriches detections with MITRE ATT&CK framework mapping. Each detection maps to specific tactics (Initial Access, Lateral Movement, Exfiltration) and techniques (T1566 Phishing, T1021 Remote Services, T1048 Exfiltration Over Alternative Protocol).

Detection Capabilities:

  • Lateral Movement Detection: Identifies when an attacker pivots between VMs using legitimate protocols (RDP, SSH, WMI, SMB). Detects anomalous access patterns even when credentials are valid.
  • Command & Control (C2) Detection: Identifies beaconing patterns (periodic outbound connections to external hosts), DNS tunneling, covert channels over HTTPS.
  • Data Exfiltration Detection: Detects unusual data volumes leaving high-value VMs, staging behavior (data aggregation before exfiltration), encoding/compression of outbound data.
  • Credential Theft Detection: Identifies Kerberoasting, pass-the-hash, and other credential harvesting techniques via network behavior patterns.

NDR Campaigns & Response

Campaign View: NDR groups related detections into "campaigns" representing a single attack lifecycle. Each campaign shows: affected VMs, timeline of events, MITRE ATT&CK mapping, confidence score (0-100%), recommended response actions.

Confidence Scoring: NDR uses ML to assign confidence scores. Low confidence (< 30%): likely benign anomaly. Medium (30-70%): requires analyst review. High (> 70%): likely malicious, recommend immediate action. Scores improve over time as the ML model learns the environment.

Integration Points:

  • SIEM Integration: Forward NDR campaigns to Splunk, QRadar, Sentinel via syslog/API.
  • SOAR Integration: Trigger automated playbooks (quarantine VM, block IP, notify SOC) based on campaign confidence score.
  • VCF Operations: NDR alerts appear in unified VCF Operations dashboard alongside performance and availability alerts.

NDR Baseline Period: 7-14 days for initial ML model training. During baseline, expect elevated false positive rate (> 10%). After training, false positive rate should drop below 2%. Plan phased rollout: start with a canary group of VMs, validate detection quality, then expand organization-wide.

Privacy and connectivity: NDR analytics stay on the SSP. Connected mode shares malware samples and telemetry with Broadcom cloud services (sandboxing, reputation, threat intelligence); Disconnected mode is documented for air-gapped deployments (SSP 5.2: Activate Network Detection and Response in Disconnected Mode for Air-Gapped Deployment) — cloud sandboxing and data sharing are unavailable there, the on-prem detectors and correlation still run. IDS/IPS and DFW enforce on hosts and Edges regardless.

Network Traffic Analysis (NTA) & Micro-Segmentation#

Two vDefend analytics features are often confused, and the 6V0-21.25 exam guide tests them as separate topics. vDefend Security Intelligence (topic 06, 'Planning Application Segmentation with vDefend Security Intelligence'; formerly NSX Intelligence, hosted on SSP/NAPP) collects DFW flow data, visualises application communication and recommends groups and rules — it answers "who talks to whom, on which ports, and how often?" and is the foundation for micro-segmentation planning. Network Traffic Analysis (NTA) is the ML anomaly-detection engine (beaconing, exfiltration, lateral movement, DGA) whose detections feed NDR campaigns (topic 17, 'NTA and NDR: proactive threat detection and response'). The planning workflow below is Security Intelligence work.

Security Intelligence Architecture

Flow Collection: NSX distributed firewall captures flow records (NetFlow/IPFIX-style) from every vNIC. Each flow record contains: source VM, destination VM, protocol, port, byte count, packet count, timestamps. Flow data is collected by the Security Intelligence service on the SSP (vDefend Firewall with ATP licence) and presented in the NSX Manager UI (Plan & Troubleshoot); flows can also be exported to VCF Operations for Networks or an external collector.

Flow Visualization: Security Intelligence presents communication patterns as a flow graph showing VMs as nodes and connections as edges. Administrators can filter by: segment, security group, application tier, time range. This visualization reveals unexpected communication paths — a web server talking directly to a database server that should route through the app tier.

Security Intelligence for Micro-Segmentation Planning

The crawl-walk-run methodology for zero-trust implementation depends on Security Intelligence (vDefend DFW 1-2-3-4 is a separate four-stage prescriptive journey — Assessment, Secure Infrastructure, Secure Environments, Secure Applications — that also runs on SSP):

Step 1 — Discover (Security Intelligence baseline, 2-4 weeks):

  • Enable Security Intelligence flow collection across all segments (clusters).
  • Allow all traffic (default allow rule) during discovery.
  • Security Intelligence builds a communication matrix: which VMs talk to which VMs, on which ports.
  • Identify application tiers automatically: web frontends (HTTP/HTTPS inbound), app servers (custom ports), databases (3306/5432/1433), management (SSH/RDP/SNMP).

Step 2 — Analyze (1-2 weeks):

  • Review the flow matrix and the recommended groups/rules in Security Intelligence.
  • Classify flows: expected (application-required), unexpected (potential security risk), management (admin access).
  • Design micro-segmentation rules based on OBSERVED traffic, not assumed traffic.
  • Target 95%+ coverage: rules should account for 95% of observed flows.
  • Identify flows that should NOT exist (lateral movement candidates, unencrypted sensitive data).

Step 3 — Design Rules (1-2 weeks):

  • Create security groups based on the application topology discovered by vDefend Security Intelligence.
  • Write DFW rules: explicit allow for expected flows, implicit deny for everything else.
  • Use tag-based groups for dynamic environments (VMs frequently created/destroyed).
  • Scope Applied-To narrowly (per application group, not "All").

Step 4 — Test (2-4 weeks):

  • Deploy rules in "Allow + Log" mode (not enforce yet).
  • Monitor DFW logs for legitimate traffic that would be blocked.
  • Tune rules: add missing allow rules for discovered flows.
  • Validate with Traceflow: inject test packets between tiers to verify path.

Step 5 — Enforce (ongoing):

  • Switch default rule from "Allow" to "Drop" per segment/tier.
  • Start with least critical workloads (dev/test).
  • Monitor for blocked legitimate traffic (check DFW drop logs).
  • Gradually enforce across all tiers over 4-8 weeks.

Step 6 — Continuous Monitoring:

  • NTA continues to analyze flows post-enforcement.
  • Detect policy violations: traffic that bypasses rules or new unexpected flows.
  • Alert on deviations from baseline (new communication patterns may indicate compromise or application changes).

Security Intelligence + VCF Automation: in mature environments, tags assigned by VCF Automation cloud templates place new VMs into tag-based groups, so they receive the DFW rules derived from the Security Intelligence recommendations without manual rule edits.

Recommendation Engine: vDefend Security Intelligence analyses the collected flows, discovers communication patterns, identifies unprotected traffic and suggests groups and DFW rules, which reduces the manual analysis needed for the initial micro-segmentation design.

Key Takeaways

  • Exam Focus: Security Intelligence (topic 06) is the planning engine — flow visibility, group and rule recommendations; NTA (topic 17) is anomaly detection that feeds NDR. In exam answers: (1) baseline flows with Security Intelligence first, (2) design rules by observed communication need, (3) reduce rule count with security groups, (4) use NTA/NDR for post-enforcement threat detection, not for rule design.

Malware Prevention & Zero-Trust Design#

Malware Prevention Module: Cloud-based and on-premises file inspection integrated with the vDefend platform. Malware Prevention uses a multi-technique approach combining machine learning, static analysis, dynamic analysis (sandboxing), and memory analysis.

Malware Prevention Architecture

Guest Introspection (GI): A lightweight agent (thin agent) runs inside each VM, monitoring file system, process, and registry activity. GI communicates with the Malware Prevention service VM (SVM) on each ESXi host of an activated cluster (VCF 9.0.1 docs p.2728). The SVM performs initial file analysis and forwards suspicious files to the cloud sandbox if needed.

Detection Pipeline:

  1. File Access Event: GI agent detects file write/execute in guest VM.
  2. Local Cache Check: SVM checks file hash against local cache of known-good/known-bad files.
  3. Static Analysis: If not cached, SVM performs static analysis (file structure, embedded strings, entropy analysis, PE header inspection).
  4. Cloud Sandbox: If static analysis is inconclusive, file is sent to Broadcom cloud sandbox for dynamic analysis (execution in isolated environment, behavior monitoring, network call analysis).
  5. Verdict: File classified as Benign, Suspicious or Malicious. Verdict cached locally for future lookups.
  6. Action: Based on policy — Allow, Block, or Quarantine.

Detection Techniques:

  • Machine Learning: Trained models classify files based on structural features without executing them. Fast (milliseconds) but may miss novel malware.
  • Static Analysis: Examines file without execution. Detects known packing, obfuscation, suspicious imports. Catches known malware variants.
  • Dynamic Analysis (Sandbox): Executes file in cloud sandbox. Monitors behavior: file system changes, registry modifications, network connections, process spawning. Catches zero-day malware that evades static analysis.
  • Memory Analysis: Inspects process memory for injection techniques, reflective DLL loading, shellcode patterns.

Deployment Considerations:

  • Service VM: one Malware Prevention SVM per ESXi host in each activated cluster (host-based service deployment, VCF 9.0.1 docs p.2728); size it per the vDefend documentation.
  • Network: SVM requires connectivity to Broadcom cloud for sandbox analysis. Latency to cloud affects time-to-verdict for unknown files.
  • Air-Gapped: Local analysis only (no cloud sandbox). Reduced detection capability for zero-day threats. ML and static analysis still functional.

Zero-Trust Security Design for VCF

Zero-trust operates on the principle "never trust, always verify." In a VCF environment, zero-trust means every network connection is authenticated, authorized, and encrypted — regardless of whether it originates inside or outside the datacenter perimeter.

Zero-Trust Architecture Pillars in VCF:

  1. Micro-Segmentation (DFW): Isolate every workload. Default deny between all VMs. Explicit allow only for required communication paths.
  2. Identity-Based Access: NSX security groups based on VM identity (tags, attributes), not IP addresses. IPs change; identity persists.
  3. Least Privilege: Each application tier gets minimum required network access. Web tier → App tier on specific ports only. No direct Web → DB.
  4. Continuous Verification: NTA monitors for deviations from approved communication patterns. IDS/IPS inspects allowed traffic for threats. NDR detects behavioral anomalies.
  5. Encryption: vSAN data-at-rest and data-in-transit encryption (VCF 9.0.1 docs p.1652), encrypted vMotion, and TLS for application traffic. NSX IPsec VPN protects site-to-site links at the Edge, not host-to-host east-west traffic.

Zero-Trust Implementation Playbook (Crawl-Walk-Run):

Crawl Phase (Weeks 1-4): Deploy DFW in monitor mode. Enable NTA flow collection. Run IDS in detection mode. Establish baseline. Zero enforcement — observe and learn.

Walk Phase (Weeks 5-12): Implement zone-based segmentation (Environment category rules). Separate Dev/Staging/Prod at network level. Deploy IDS/IPS signatures with tuned profiles. Begin application-tier micro-segmentation on non-critical workloads.

Run Phase (Weeks 13-20): Full micro-segmentation on all workloads. Default deny rule enforced. IPS in prevention mode. NDR active with automated response playbooks. Continuous NTA monitoring for policy compliance.

Common Failure Patterns:

  • Jumping to "Run" without baseline → production outages from blocked legitimate traffic.
  • Using IP-based rules instead of tag-based groups → rules break when VMs change IP.
  • Enabling all IDS signatures simultaneously → alert fatigue from false positives.
  • Not involving application teams → missing required communication paths in rules.

Key Takeaways

  • Exam Focus - Zero-Trust Implementation: the pattern is crawl-walk-run. Organisations fail by jumping to full enforcement without a baseline and cause production outages. Correct approach: (1) baseline flows with Security Intelligence for about 30 days, (2) enforce tier by tier starting with the least complex application, (3) keep a default allow-with-logging rule until the matrix is validated, then flip the default to deny per application, (4) monitor continuously and treat new flows as change requests.

vDefend Suite Architecture — All Components Deep Dive#

vDefend represents Broadcom's consolidated security suite for VMware Cloud Foundation. It unifies multiple formerly separate security products into a single platform with integrated licensing and management.

vDefend Product Consolidation

vDefend replaces and consolidates:

  • NSX Distributed Firewall (DFW) → vDefend Distributed Firewall
  • NSX Gateway Firewall (GWF) → vDefend Gateway Firewall
  • NSX Advanced Threat Prevention (ATP) → vDefend Advanced Threat Prevention (includes IDS/IPS, Malware Prevention, NTA, NDR)

All vDefend components run natively within the NSX infrastructure — no additional appliances required for DFW, Gateway FW, or IDS/IPS. Malware Prevention requires Guest Introspection and a service VM (SVM) on each host of an activated cluster. Security Intelligence, NDR and Malware Prevention run on the Security Services Platform (SSP), deployed separately and onboarded to NSX Manager.

Component Integration Architecture:

┌─────────────────────────────────────────────────┐
│                vDefend Platform                   │
│  ┌──────────────┐  ┌──────────────┐             │
│  │ DFW          │  │ Gateway FW   │  Layer 2-7   │
│  │ (E-W traffic)│  │ (N-S traffic)│  Firewalling │
│  └──────────────┘  └──────────────┘             │
│  ┌──────────────┐  ┌──────────────┐             │
│  │ IDS/IPS      │  │ Malware Prev │  Threat      │
│  │ (Signatures) │  │ (Sandbox+ML) │  Prevention  │
│  └──────────────┘  └──────────────┘             │
│  ┌──────────────┐  ┌──────────────┐             │
│  │ NTA          │  │ NDR          │  Detection & │
│  │ (Flow Analy.)│  │ (ML Behavior)│  Response    │
│  └──────────────┘  └──────────────┘             │
└─────────────────────────────────────────────────┘

Data Flow Between Components:

  1. DFW/GWF → generates firewall logs (allow/deny/drop events)
  2. IDS/IPS → generates signature-match alerts
  3. Malware Prevention → generates file verdict events
  4. NTA → generates flow baseline and anomaly alerts
  5. NDR Aggregation Engine → correlates all signals into unified campaigns

All components feed telemetry to NSX Manager, which provides a unified security dashboard. Security Intelligence, NDR and Malware Prevention analytics run on the SSP; Connected mode adds Broadcom cloud sandboxing and threat intelligence, Disconnected mode keeps everything on premises.

Licensing Tiers

vDefend Firewall: Includes DFW + Gateway Firewall. Sold as a per-core add-on to VCF; it is not included in the VCF primary licence (VCF 9.0.1 docs p.1458 primary licences; p.1151 "Firewall Add-On is required").

vDefend Firewall with Advanced Threat Prevention: Includes DFW + Gateway FW + IDS/IPS + Malware Prevention + NTA + NDR. Licensed per core as an add-on; check the current Broadcom vDefend SKU list for whether ATP is bought as a combined Firewall-with-ATP SKU or added on top of vDefend Firewall.

Core Counting: vDefend licenses count physical CPU cores on ESXi hosts where vDefend features are activated. Minimum 16 cores per CPU (same as VCF base licensing).

Operational Workflow — Day 2 Security Operations

Security Operations Center (SOC) Integration:

  • Tier 1 (Alert Triage): VCF Operations dashboards show DFW allow/deny trends, IDS alert volume, NDR campaign count. SOC analyst reviews high-confidence NDR campaigns first.
  • Tier 2 (Investigation): Analyst uses Traceflow to validate alert, reviews NTA flow history for affected VMs, checks Malware Prevention verdicts for related files.
  • Tier 3 (Response): Remediate via DFW Emergency rules (quarantine compromised VM), update IDS signatures, submit false-positive reports, update micro-segmentation rules.

Compliance Reporting: vDefend provides audit-ready reports for: DFW rule changes (who changed what, when), IDS/IPS detection history, flow analysis compliance (which VMs communicate across compliance boundaries). Maps to frameworks: PCI-DSS (network segmentation), HIPAA (access controls), SOC 2 (monitoring).

DFW Rule Engine Internals#

The Distributed Firewall rule engine is the enforcement engine for zero-trust security. Understanding its internals is critical for performance optimization and troubleshooting.

Rule Processing Pipeline

When a packet arrives at a VM's vNIC, the VSIP filter processes it through this pipeline:

  1. Connection Table Lookup: Check if packet belongs to an existing tracked connection. If yes, apply the existing connection's action (fast path — bypasses rule evaluation). If no, proceed to rule evaluation (slow path).
  1. Rule Evaluation (Slow Path): Evaluate rules category by category, left to right: Ethernet → Emergency → Infrastructure → Environment → Application (the default allow rule is the last rule of the Application category). Within each category, evaluate sections top-to-bottom. Within each section, evaluate rules top-to-bottom. First matching rule wins — no further evaluation.
  1. Action Execution: Apply the matched rule's action: Allow (pass packet, create connection table entry), Drop (silently discard packet), Reject (discard + send TCP RST or ICMP unreachable to source).
  1. Logging: If rule has logging enabled, write syslog entry with: rule ID, action, source/dest IP, port, protocol, timestamp.

Connection Table Architecture

The connection table is the performance-critical component:

  • Size: finite per host; check the vDefend configuration maximums for the current value.
  • Entry Lifecycle: Created on first packet of new connection. Updated with byte/packet counters. Removed on connection close (TCP FIN/RST) or timeout, per the Firewall Session Timer profile (check the vDefend documentation for the default TCP/UDP/ICMP values).
  • Monitoring: esxtop does not show DFW connection counts; use nsxcli on the host for per-filter rule and flow counts, or NSX Manager / VCF Operations.
  • Saturation: When the table is full, new connections are DROPPED. Symptom: intermittent connectivity for new connections while existing connections work fine.
  • Tuning: Increase table size for high-connection hosts (load balancers, proxies). Decrease timeouts for environments with many short-lived connections.

Applied-To Optimization

Applied-To controls which hosts/vNICs evaluate a rule. This is the single most important performance lever:

Scope: "DFW" (All): Every vNIC on every host evaluates this rule. Worst performance. Use only for Emergency rules (quarantine).

Scope: Security Group: Only vNICs in the group evaluate. Recommended. Example: "Applied-To: Web-Tier-Group" — only web VMs process the rule.

Scope: Specific VM: Single VM evaluates. Most restrictive. Use for targeted rules.

Performance Impact: a new connection is evaluated only against the rule table of the vNIC filter it passes through, not every vNIC on the host. With Applied-To = DFW, every vNIC filter on every host carries all 1,000 rules (memory and realisation cost, and a longer per-vNIC lookup); scoping a rule to a 10-VM group means only those 10 filters carry it. Monitor per-filter rule and flow counts with nsxcli on the host.

Rule Consolidation Strategies

  1. Use Security Groups instead of individual IPs: One rule with group source/destination instead of N rules per VM pair.
  2. Use service groups: Combine related ports (HTTP + HTTPS + HTTP/2) into a single service object.
  3. Merge bi-directional rules: If A→B and B→A on same ports, use a single "allow" rule (stateful tracking handles return traffic automatically).
  4. Eliminate duplicate/shadow rules: Use Firewall Rule Analysis (SSP) and rule hit counts to identify and remove them.

Troubleshooting DFW Issues

Packet drops with no matching rule: the DFW has NO implicit deny. The last rule in the Application category is the default rule, set to ALLOW out of the box so VM-to-VM traffic is not broken during host preparation; it is configurable to Drop/Reject under Security > Distributed Firewall and cannot be deleted. If traffic is dropping with no matching rule, someone has already flipped that default to Drop — check it first.

Intermittent connectivity: Check connection table utilization. If near capacity, existing connections work but new ones fail randomly.

Rule not matching expected traffic: Verify group membership is correct (dynamic groups may have stale membership). Use Traceflow to see which rule matched. Check rule order — a broader rule higher in the list may be matching first.

High CPU on ESXi host: Check DFW rule count per vNIC (nsxcli > get firewall rules). Reduce Applied-To scope. Consider reducing IDS/IPS signature count if combined with DFW.

DFW Exclusion List Issues: VMs on the exclusion list bypass ALL DFW rules. If a VM stops being firewalled unexpectedly, check the exclusion list first.

Key Takeaways

  • Exam Tip: for DFW performance bottlenecks consider (1) rule count per vNIC, (2) Applied-To scope (avoid "DFW"/All), (3) connection table saturation, (4) group-membership explosion (dynamic groups with thousands of members). Solutions: narrow Applied-To, consolidate rules, use static groups where membership is stable, and review with Firewall Rule Analysis and rule hit counts.

Zero-Trust Migration Playbook#

Transitioning from perimeter-based to zero-trust requires phased deployment. This playbook provides a practical migration path for VCF environments using vDefend.

Pre-Migration Assessment (Week 0)

  1. Inventory: Catalog all VMs, applications, and network segments in the VCF environment. Tag each VM with application role, tier, environment, and compliance zone.
  2. Communication Mapping: Use Security Intelligence to discover existing communication patterns (2-4 week baseline).
  3. Risk Assessment: Identify high-value assets (databases with PII, financial systems, intellectual property). These get micro-segmented first.
  4. Stakeholder Alignment: Brief application teams on the migration plan. They must validate that discovered communication paths are correct and identify any seasonal or batch traffic patterns.

Phase 1 — Zone Segmentation (Weeks 1-4)

Objective: Separate major zones (Dev/Staging/Prod/Management) with DFW rules.

Implementation:

  • Create environment tags: env:dev, env:staging, env:prod, env:mgmt.
  • Apply tags to all VMs (manual or via VCF Automation integration).
  • Create DFW rules in Environment category:
    • Allow intra-zone traffic (dev→dev, prod→prod).
    • Deny cross-zone traffic by default (dev→prod blocked).
    • Explicit exceptions for required cross-zone communication (CI/CD pipeline, monitoring).
  • Deploy in monitor+log mode for 1 week, then enforce.

Validation: Traceflow between zones should show deny. VCF Operations alerts should show zero blocked legitimate traffic.

Phase 2 — Application Tier Micro-Segmentation (Weeks 5-12)

Objective: Isolate application tiers (Web, App, DB) within each zone.

Implementation:

  • Create application tier tags: tier:web, tier:app, tier:db, tier:mgmt-tools.
  • Design tier-based rules in Application category:
    • Web→App: Allow on application ports (TCP 8080, 8443).
    • App→DB: Allow on database ports (TCP 3306, 5432, 1433).
    • Web→DB: DENY (enforce separation of concerns).
    • Mgmt→All: Allow on management ports (SSH 22, HTTPS 443, SNMP 161).
    • All→DNS/NTP: Allow infrastructure services.
  • Enable IDS/IPS in detection mode on inter-tier traffic.
  • Deploy per-application, starting with least critical.

Validation: Traceflow validates allowed paths. Security Intelligence confirms no unexpected flows. IDS/IPS generates baseline alert volume.

Phase 3 — Workload-Level Micro-Segmentation (Weeks 13-20)

Objective: Isolate individual workloads — each VM group gets only the access it needs.

Implementation:

  • Create application-specific groups: app:payroll-web, app:payroll-api, app:payroll-db.
  • Design workload-level rules:
    • payroll-web→payroll-api: Allow TCP 8443.
    • payroll-api→payroll-db: Allow TCP 5432.
    • payroll-*→anything-else: DENY.
  • Switch default rule from "Allow" to "Drop" for each secured application.
  • Enable IPS (prevention mode) after 30-day IDS tuning.
  • Enable Malware Prevention on high-value VMs.

Phase 4 — Continuous Enforcement (Ongoing)

  • Security Intelligence monitors for new/changed communication patterns; NTA detects behavioural anomalies.
  • NDR detects behavioral anomalies and lateral movement.
  • Automated response: High-confidence NDR campaigns trigger DFW Emergency quarantine rules via SOAR integration.
  • Rule hygiene: Monthly review using Firewall Rule Analysis (resolve duplication, redundancy, shadowing and consolidation findings) and DFW rule hit counts (remove unused rules).
  • Compliance reporting: Generate quarterly audit reports mapping DFW rules to compliance requirements (PCI-DSS, HIPAA).

Migration Timeline Summary:

  • Small environment (< 200 VMs): 8-12 weeks
  • Medium environment (200-1000 VMs): 12-20 weeks
  • Large environment (> 1000 VMs): 20-30 weeks with dedicated team

Common Pitfalls:

  1. Skipping the Security Intelligence flow baseline → designing rules based on assumptions, not facts.
  2. Enforcing too quickly → production outages from blocked legitimate traffic.
  3. IP-based rules → rules break when VMs migrate or change IP.
  4. Ignoring application team input → missing seasonal or batch communication paths.
  5. Not monitoring after enforcement → security rules become stale as applications evolve.

IDS/IPS Design & Tuning#

Intrusion Detection/Prevention is a critical layer of vDefend, providing signature-based threat detection for both east-west and north-south traffic.

IDS/IPS Design for VCF Environments

Deployment Strategy — Defense in Depth:

Layer 1 — Gateway IDS/IPS (Perimeter):

  • Deploy on Tier-0 gateway rules for all inbound/outbound traffic.
  • Profile: Critical + High severity, widened to Medium for internet-exposed gateways once tuned.
  • Purpose: Catch known threats at the boundary before they reach workloads.
  • Capacity: Edge node sizing affects throughput. Large Edge VM (8 vCPU, 32GB) handles up to 10 Gbps with IDS enabled.

Layer 2 — Distributed IDS/IPS (East-West):

  • Deploy on DFW rules between application tiers.
  • Profile: Critical + High severity (avoid broad all-severity profiles on high-traffic tiers).
  • Purpose: Detect lateral movement — attacker already inside, moving between VMs.
  • Capacity: Per-host processing. Monitor ESXi CPU impact.

Layer 3 — Application-Specific:

  • Deploy custom signature profiles per application tier.
  • Web tier: HTTP/TLS inspection signatures (SQL injection, XSS, path traversal).
  • Database tier: SQL protocol anomaly signatures.
  • Infrastructure: SSH brute-force, SNMP exploitation, management protocol abuse.

Signature Profile Design

Tiered Approach — Match profile severity to workload criticality:

Production Database Tier:

  • Profile: Critical + High severity, filtered by product affected to the database platform, plus custom SQL signatures.
  • Mode: IPS (prevention) after tuning.
  • Rationale: High-value data, low tolerance for compromise, narrow protocol set (easy to tune).

Production Web Tier:

  • Profile: Critical + High severity, filtered by product affected to the web/application stack.
  • Mode: IPS after 45-day tuning.
  • Rationale: High attack surface (internet-facing), diverse traffic patterns require longer tuning.

Development Environment:

  • Profile: Critical severity only.
  • Mode: IDS (detect-only) permanently.
  • Rationale: Frequent code changes generate false positives. Detection provides visibility without blocking developer productivity.

Management Infrastructure (vCenter, NSX, ESXi):

  • Profile: broad (Critical through Low severity), filtered by product affected to the management products in use.
  • Mode: IPS.
  • Rationale: Management plane compromise is catastrophic. Narrow, predictable traffic patterns produce few false positives.

Tuning Deep Dive

False Positive Categories:

  1. Legitimate scanning tools (vulnerability scanners, SNMP monitors) → Whitelist source IP + signature ID.
  2. Legacy protocols (older TLS versions, non-standard HTTP) → Disable specific protocol anomaly signatures for affected segments.
  3. Custom applications using unusual patterns (binary protocols over HTTP ports) → Create exclusion rules or custom allow signatures.
  4. Backup/replication traffic (high-volume data transfer triggers DDoS signatures) → Whitelist backup server IPs for DDoS signature category.

Metrics to Track:

  • True Positive Rate: % of alerts that are genuine threats. Target: > 80%.
  • False Positive Rate: % of alerts that are benign. Target: < 5% after tuning.
  • Mean Time to Detect (MTTD): Time from attack start to first alert. Target: < 5 minutes for known signatures.
  • Signature Coverage: % of MITRE ATT&CK techniques covered. Track gaps quarterly.

Operational Procedures:

  • Signature Update Review: Monthly review of new signatures before activation. Test in lab/dev first.
  • Quarterly Tuning Review: Re-evaluate disabled signatures, check for new false positive patterns.
  • Incident Response Integration: IDS/IPS alerts feed into SIEM/SOAR. Define playbooks for high-severity signatures (CVSS 9+): auto-quarantine → SOC notification → forensic capture → remediation.

Key Takeaways

  • Exam Tip: for IDS/IPS at scale mention (1) Gateway IDS/IPS for the perimeter, (2) Distributed IDS/IPS for east-west, (3) profile selection per tier (signature severity, CVSS, product affinity), (4) the false-positive tuning process, (5) custom signatures for zero-days. Do not answer "enable IPS on all rules" — that ignores Edge/host throughput cost, alert volume and the need for per-tier profiles.

Network Detection & Response (NDR) Deep Dive#

Network Detection & Response is the vDefend analytics service that runs on the Security Services Platform (SSP 5.x; NSX Application Platform before it) and is activated from SSP in Connected or Disconnected mode. NDR provides behavioral threat detection using machine learning, complementing signature-based IDS/IPS detection.

NDR Detection Engine Architecture

The NDR detection engine uses three types of ML models:

Supervised Models: Trained on known-malicious traffic patterns (labeled datasets from Broadcom threat research). Detect known attack variants even when signatures don't exist. Updated continuously via cloud model updates.

Unsupervised Models: Learn "normal" for each environment without labeled data. Detect deviations from learned baseline. Catches zero-day attacks that don't match any known pattern. Requires 7-14 day baseline period per environment.

Deep Learning Models: Neural networks analyzing complex traffic patterns across multiple dimensions simultaneously. Detect subtle, multi-stage attacks that simpler models miss. Higher computational cost but lower false positive rate for complex scenarios.

MITRE ATT&CK Integration

Every NDR detection is mapped to the MITRE ATT&CK framework:

Initial Access (TA0001): Phishing links in email attachments, exploitation of public-facing applications.

Execution (TA0002): Remote code execution, scripting engine abuse.

Persistence (TA0003): Scheduled tasks, registry modifications for auto-start.

Lateral Movement (TA0008): RDP/SSH pivoting, SMB file sharing exploitation, WMI remote execution.

Collection (TA0009): Data staging on network shares, screen capture uploads.

Exfiltration (TA0010): Data transfer to cloud storage, DNS tunneling, HTTP/S data smuggling.

Command & Control (TA0011): Beaconing patterns, domain generation algorithms (DGA), encrypted C2 channels.

This mapping enables SOC analysts to understand attack progression and prioritize response.

Campaign Correlation Engine

NDR's primary value is correlating individual low-confidence events into high-confidence campaigns:

Example Attack Campaign:

T+0: NTA detects unusual DNS query from VM-web-01 to suspicious domain (confidence: 15%).

T+5min: IDS detects encoded PowerShell download on VM-web-01 (confidence: 40%).

T+15min: NDR detects VM-web-01 making RDP connection to VM-db-01 — first ever RDP between these VMs (confidence: 55%).

T+20min: NDR detects large data transfer from VM-db-01 to external IP (confidence: 70%).

Individual events: Low confidence, easily lost in noise.
Correlated Campaign: High confidence (85%), clear attack chain, automated MITRE mapping.

Response Actions:

  • Automated: DFW Emergency rule quarantines VM-web-01 (block all traffic).
  • SOC Alert: Campaign details pushed to SIEM with full timeline.
  • Evidence Capture: NTA flow records preserved for forensic analysis.
  • Playbook Execution: SOAR platform runs incident response procedures.

NDR Deployment Best Practices

  1. Start Small: Deploy on one workload domain. Validate detection quality before expanding.
  2. Tune During Baseline: Work with the SOC team during the 14-day baseline to identify expected anomalies (batch jobs, backup windows, maintenance activities).
  3. Integrate Early: Connect NDR to SIEM/SOAR before going live. Ensure automated playbooks are tested.
  4. Decide the mode up front: Connected mode needs outbound connectivity from the SSP to Broadcom cloud services (sandboxing, threat intelligence, data sharing); Disconnected mode is for air-gapped estates and forgoes those. Monitor SSP health and the NSX-to-SSP integration, not an internet link.
  5. Review Campaigns Weekly: SOC analysts should review all medium+ confidence campaigns weekly. Feedback loop (confirm/dismiss) improves ML model accuracy.

Limitations:

  • Needs the Security Services Platform (sizing, Kubernetes footprint, lifecycle) in addition to NSX; cloud sandboxing and data sharing are unavailable in Disconnected mode.
  • Encrypted payload inspection is NOT performed (only metadata analysis). End-to-end encrypted C2 channels may evade if they mimic normal traffic patterns perfectly.
  • 7-14 day baseline means no meaningful detections during initial deployment period.

Key Takeaways

  • Exam Tip: when designing NDR deployment, emphasise (1) analytics on the on-prem SSP with Connected or Disconnected (air-gapped) mode, (2) correlation across flows, hosts and time rather than single-packet detection, (3) MITRE ATT&CK mapping for standardised response, (4) SIEM/SOAR integration for automated playbooks, (5) confidence scoring and analyst feedback to suppress noise, (6) a 7-14 day baseline before expecting meaningful campaigns.

Lab Exercise: Zero-Trust Enforcement with vDefend#

Lab Exercise: Zero-Trust Enforcement with vDefend — Practical Reference

This exercise guides you through implementing a complete zero-trust micro-segmentation deployment in a Holodeck lab environment using vDefend DFW, IDS/IPS, and Security Intelligence.

Lab Prerequisites:

  • VCF management domain operational with NSX configured.
  • At least 6 VMs deployed across 3 application tiers: 2x Web (nginx), 2x App (tomcat), 2x DB (postgresql).
  • All VMs connected to NSX overlay segments.
  • VCF Operations deployed and collecting metrics.

Exercise 1 — Security Intelligence Flow Baseline (30 minutes):

  1. Access NSX Manager → Plan & Troubleshoot (Security Intelligence views; requires the SSP/NAPP service onboarded).
  2. Enable flow collection for all workload segments.
  3. Generate traffic: curl from Web VMs to App VMs (port 8080), App VMs query DB VMs (port 5432).
  4. Review flow visualization — verify expected communication paths appear.
  5. Identify unexpected flows (e.g., Web→DB direct communication that should not exist).
  6. Export flow data for rule design.

Exercise 2 — Security Group Design (20 minutes):

  1. NSX Manager → Inventory → Groups → Add Group.
  2. Create dynamic groups:
  • "Web-Tier": Membership criteria = Tag equals "tier:web"
  • "App-Tier": Membership criteria = Tag equals "tier:app"
  • "DB-Tier": Membership criteria = Tag equals "tier:db"
  1. Apply tags to VMs: NSX Manager → Inventory → Virtual Machines → select VM → Tags → Add.
  2. Verify group membership: click each group → Members tab → confirm correct VMs listed.

Exercise 3 — DFW Rule Implementation (30 minutes):

  1. NSX Manager → Security → Distributed Firewall.
  2. Create new section "App-Microseg" in Application category.
  3. Add rules:

Rule 1: Source=Web-Tier, Dest=App-Tier, Service=TCP-8080, Action=Allow, Applied-To=Web-Tier,App-Tier, Log=Yes

Rule 2: Source=App-Tier, Dest=DB-Tier, Service=TCP-5432, Action=Allow, Applied-To=App-Tier,DB-Tier, Log=Yes

Rule 3: Source=Any, Dest=Web-Tier, Service=HTTPS, Action=Allow, Applied-To=Web-Tier, Log=Yes

Rule 4: Source=Any, Dest=Any, Action=Drop, Applied-To=Web-Tier,App-Tier,DB-Tier, Log=Yes

  1. Publish rules.
  2. Validate: curl from Web VM to App VM on 8080 → should succeed.
  3. Validate: curl from Web VM to DB VM on 5432 → should be BLOCKED.
  4. Use Traceflow: NSX Manager → Plan & Troubleshoot → Traceflow → trace Web→DB on 5432 → verify DROP by rule 4.

Exercise 4 — IDS/IPS Deployment (20 minutes):

  1. NSX Manager → Security → IDS/IPS & Malware Prevention → Settings → Enable.
  2. Download latest signature bundle.
  3. Create IDS profile: "Lab-Detection" with Critical + High signatures enabled.
  4. Apply IDS profile to rule 1 (Web→App traffic) and rule 2 (App→DB traffic).
  5. Generate test alert: use curl to send a test pattern matching a known signature (e.g., SQL injection attempt in URL parameter).
  6. Verify alert appears in NSX Manager → Security → IDS/IPS → Events.

Exercise 5 — Validation & Monitoring (20 minutes):

  1. VCF Operations → dashboards → verify DFW metrics (allow/drop counts, top talkers).
  2. Check syslog for DFW rule hit logs — verify expected allow/drop entries.
  3. Review IDS events — confirm test alert captured with correct signature details.
  4. Generate a compliance report from Security Intelligence showing micro-segmentation coverage.

Lab Completion Criteria:

  • DFW rules enforce tier-based isolation (Web↔App↔DB allowed, Web→DB blocked).
  • IDS detects simulated attack on allowed traffic paths.
  • Traceflow confirms correct rule matching for both allowed and denied paths.
  • VCF Operations dashboard shows DFW and IDS metrics.

Security Services Platform (SSP) and NSX Application Platform (NAPP)#

Where the advanced vDefend services run.

vDefend Distributed Firewall and Gateway Firewall need nothing beyond NSX Manager and the transport nodes. The analytics-heavy services do: Security Intelligence, Network Traffic Analysis (NTA), Malware Prevention and Network Detection and Response (NDR) are hosted on a Kubernetes-based platform. In the VCF 5.2 / NSX 4.2 line that platform is the NSX Application Platform (NAPP); in VCF 9.x it is the Security Services Platform (SSP, 5.x line), deployed by the SSP Installer appliance.

NAPP (VVS 1.0 for VCF 5.2):

  • Runs on a dedicated Tanzu Kubernetes Grid cluster managed by a vSphere Supervisor; the Supervisor of the management domain is recommended for all workload-domain NAPP instances.
  • One NAPP instance per NSX Manager cluster (1:1); the NAPP Automation Appliance supports up to five instances.
  • Sizing: 3 control-plane nodes (2 vCPU / 4 GB / 64 GB) and at least 4 worker nodes (16 vCPU / 64 GB / 64 GB), 100 % reserved in production; deploy minimal, run Metrics for 7 days, size with the Broadcom sizing tool (KB 373793), scale out, then activate Security Intelligence.
  • Prerequisites are design decisions: NAPP management network on the VCF VM management network (SEC-NAPP-008), seven reserved IPs (SEC-NAPP-009), a front-end and a workload network (SEC-NAPP-010/-011), two FQDNs (SEC-NAPP-012), HA-Proxy as the only load balancer, in the 3-NIC front-end option (SEC-NAPP-013/-014).

SSP (VVS 9.X for VCF 9.0):

  • SSP Installer VMs and SSP instances live in the management domain primary cluster or a dedicated cluster (SEC-SSP-CFG-001); one Installer per VCF domain (SEC-SSP-CFG-002).
  • Availability: vSphere HA plus SSP's own Machine Health Check (MHC) that redeploys unhealthy nodes, so no manual DRS anti-affinity rules (SEC-SSP-CFG-005); in multi-AZ designs MHC is disabled and vSphere HA alone restarts the nodes in the second zone (SEC-SSP-CFG-007).
  • Network: a dedicated DVPG, NSX segment or VPC Public subnet with 23 IP addresses per instance at maximum scale (SEC-SSP-NET-002); the Installer needs the vCenter root CA certificate and runs a pre-check before deployment.
  • NDR on SSP runs in Connected mode (outbound 443 to the vDefend ATP Cloud Service) or Disconnected mode for air-gapped sites.

Licensing: Security Intelligence ships with vDefend Firewall; IDS/IPS, Malware Prevention, NTA and NDR need vDefend Firewall with Advanced Threat Prevention (ATP). IDS/IPS itself does not need NAPP/SSP.

Sources: VVS 1.0 pp.867-874; VVS 9.X Lateral Security (SEC-SSP-*); VCF 9.0.1 NSX Application Platform support bundle (features hosted).

Key Takeaways

  • DFW/GFW need only NSX; Security Intelligence, NTA, Malware Prevention and NDR need NAPP (5.2) or SSP (9.x)
  • NAPP: 1:1 with an NSX Manager cluster, 3 control-plane + ≥4 worker nodes, size from 7 days of Metrics
  • SSP: management-domain placement, MHC instead of anti-affinity rules, MHC off for multi-AZ, 23 IPs per instance

Planning Application Segmentation with Security Intelligence#

Security Intelligence (formerly NSX Intelligence) is the vDefend Firewall feature on NAPP/SSP that visualises groups, VMs, physical servers and the flows between them, and recommends groups and DFW rules from observed traffic.

When to use it: where vendor documentation of an application's flows exists, build rules from it; for custom or undocumented applications, let Security Intelligence analyse flows over time and recommend rules (VVS 1.0, Security Intelligence in the VI Workload Domain).

The recommendation inputs are design decisions (VVS 1.0 Table 652):

  • SEC-VIWLD-016: use Security Intelligence for micro-segmentation in workload domains (requires NAPP).
  • SEC-VIWLD-017: define the scope of each application as an NSX group containing all its members.
  • SEC-VIWLD-018: choose the connectivity strategy — Allow list drops flows that match no rule; Deny list allows them.
  • SEC-VIWLD-019: choose the time range of flow collection per application (wrong range = missing or stale rules).
  • SEC-VIWLD-020: choose the traffic type considered (incoming, intra-application).
  • SEC-VIWLD-021: choose Layer-4 or Layer-7 rules — APP-IDs detected in flows can generate L7 rules.
  • SEC-VIWLD-022: Emergency, Infrastructure and Environment rules are never generated; recommendations publish into the Application category.

Workflow (NSX Manager): Plan & Troubleshoot > Discover & Take Action shows the canvas (Groups or Computes view, Flows bar); Plan & Troubleshoot > Recommendations > Start New Recommendation sets name, scope, time range (last hour / 24 hours / week / month), traffic type, connectivity strategy, group-reuse threshold and infrastructure exclusion; Start Discovery runs the job to Ready to Publish; a Monitor toggle re-checks scope and strategy hourly.

Staging enforcement: publish first with a Deny-list strategy and logging (default rule allows), review flows that hit the default rule, then rerun with Allow list (default rule drops). Logging follows SEC-VIWLD-023/-024: log granular rules, not broad rules or the default L3 rule.

Key Takeaways

  • Scope, connectivity strategy, time range, traffic type and L4/L7 are explicit decisions (SEC-VIWLD-017..021)
  • Recommendations land in the Application category only; infrastructure and environment rules stay manual
  • Allow list = default drop, Deny list = default allow — stage from Deny list to Allow list

Context-Aware Firewall (L7) and Identity Firewall#

Context-aware firewall:

NSX has two profile types, context profiles and Layer 7 access profiles. A context profile carries attributes — App ID, Custom URL, Domain (FQDN) Name, URL Category and URL Reputation — and is attached to DFW or Gateway Firewall rules. App IDs can carry sub-attributes such as TLS_Version and CIPHER_SUITE (one App ID with sub-attributes per profile; sub-attributes clear when several App IDs are combined). Both App ID and FQDN can sit in one profile. Create them under Security > Profiles > Context Profile (or Inventory > Profiles); FQDN usage is listed under Inventory > Profiles > FQDNs.

Design rule from the VVS (SEC-VIWLD-014/-015): couple every L7 rule with its correlating L4 service (HTTP context profile with TCP/80, MYSQL with TCP/3306) and create one rule per protocol and context profile; a service of Any with a context profile is allowed but sends needless packets to deep packet inspection.

Identity Firewall (IDFW):

Active Directory supplies the user identity. Register one or more Windows domains under System > Identity Firewall AD (name, NetBIOS name, Base DN such as DC=corp,DC=local, delta synchronisation interval, LDAP server, optional Event Log Server). NSX Manager synchronises groups, users and their relationships; AD changes are visible only after a delta or full sync. Identity-based groups are then used in DFW rules.

Two detection paths for log-on events: Guest Introspection / VM Tools, or event log scraping from the domain controllers (the event log reader account needs Security event log read permission; NTP must be consistent). Group-membership changes take effect only after the user logs out and in again. Windows 2008 is not supported as AD or RDSH server. Identity Firewall is part of the vDefend Firewall add-on and its state is synchronised on Stateful Active/Active gateways.

Sources: VCF 9.0.1 Context Profiles p.2226; Configuring Active Directory and Event Log Scraping pp.2663-2664; VVS 1.0 Layer 7 Policy Considerations.

Key Takeaways

  • Context profile attributes: App ID (with TLS_Version / CIPHER_SUITE sub-attributes), Custom URL, FQDN, URL Category, URL Reputation
  • Pair each L7 rule with its L4 service; one rule per protocol and profile
  • IDFW: register AD (NetBIOS, Base DN, LDAP, optional Event Log Server); changes need a sync and a user re-logon

Protecting Container Workloads (Antrea and VKS)#

Kubernetes clusters that run the Antrea CNI can be registered to NSX so that the DFW policy model extends to pods.

Integration architecture (VCF 9.0.1 pp.2573-2575): the Antrea NSX Adapter runs as a pod on a control-plane node and has two sub-components. The Management Plane Adapter talks to the NSX Policy management plane, the Kubernetes API server and the Antrea Controller: it reports the cluster inventory (pods, Ingress, Services, Network Policies, Namespaces, Nodes) to NSX Manager and answers policy-statistics queries. The Central Control Plane Adapter talks to the NSX CCP and the Kubernetes API server to realise NSX policies as Antrea policies. The inventory appears in NSX Manager and cluster health can be monitored there.

Security policy for Antrea clusters (VCF 9.0.1 pp.2578-2580):

  • Antrea groups are created in NSX and used in security policies to protect pod-to-pod traffic; separate policies secure traffic between Antrea clusters and VMs on NSX networks.
  • Only Layer 3/4 policies; rules allowed in the Emergency, Infrastructure, Environment and Application categories (Layer 2 Ethernet rules are not supported). Categories map to Antrea tiers and are enforced in that order.
  • Sources, destinations and Applied To may contain only Antrea groups; Applied To works at policy or rule level (policy level wins when both are set).
  • Services: TCP and UDP only, destination ports only; raw port/protocol combinations supported for TCP/UDP.
  • Policy and rule statistics are supported (not aggregated across clusters automatically).
  • Antrea groups are not supported under NSX projects (multi-tenancy).
  • Group membership for VM-to-cluster policies can match Ingress, Gateway, LoadBalancer or NodePortLocal Services by cluster, namespace and tag conditions.

In VCF 9.0 the vSphere Kubernetes Service (VKS) clusters use Antrea by default, so this is the path to a single policy model across VMs and containers; Container Security with Antrea is included in the vDefend Firewall add-on (VVS 1.0 licensing).

Key Takeaways

  • Antrea NSX Adapter = MP Adapter (inventory, statistics) + CCP Adapter (policy realisation)
  • Antrea policies: L3/L4 only, four categories (no Ethernet), Antrea groups only, TCP/UDP destination ports
  • Not supported under NSX projects; statistics not aggregated across clusters

Security Automation#

Why automate: the VVS makes the operational practices themselves design decisions — update the Edge Node IP group before commissioning hosts (SEC-LCM-006), create workload-domain groups and tags before deployment (SEC-LCM-007), update parent groups for shared NSX Managers (SEC-LCM-008), re-tag vCenter after a reduced downtime upgrade (SEC-LCM-009), update the Application policy per new workload domain (SEC-LCM-010). Each says the administrator must follow a runbook and/or an automated process.

Interfaces:

  • NSX Policy API (declarative, hierarchical intent: PATCH /policy/api/v1/infra/... for groups, security policies, rules). It is the supported interface; the Manager API is deprecated. Idempotent PATCH calls suit configuration-as-code and SOAR playbooks (for example adding a VM external_id to a quarantine group).
  • Terraform: the VVS 1.0 implementation uses the open-source Terraform provider to automate management-domain DFW configuration and application security in workload domains; VCF 9.0 adds Terraform configuration for VCF Automation and Terraform support for VPC objects.
  • PowerCLI: Managing the NSX Policy API with VMware Cloud Foundation PowerCLI (VCF 9.0.1 p.7954).
  • Security Intelligence recommendations publish DFW groups and rules without hand-writing them; emergency, infrastructure and environment rules remain outside (SEC-VIWLD-022).
  • Event-driven response: NDR campaign events to a SIEM, SOAR playbook calling the Policy API to populate an Emergency-category quarantine group (see labs vdefend-03 and net-virt-03).

Guardrails: keep management-domain NSX Managers in the User Excluded Group (SEC-MWLD-005) so automation mistakes cannot lock the control plane out; publish default-drop rules last with logging (SEC-MWLD-038/-043/-050); run automation under a dedicated role (Security Admin or a cloned custom role) rather than Enterprise Admin.

Key Takeaways

  • Lifecycle practices SEC-LCM-006..010 are the automation backlog: runbook or automated process
  • Policy API (declarative PATCH) and the Terraform provider are the supported automation surfaces
  • Exclusions and logged default drops are the safety net for automated policy changes

Role-Based Access Control for Security Administration#

NSX ships built-in roles, each with a fixed permission matrix for every feature area (VCF 9.0.1 Roles and Permissions): Enterprise Admin (EA), Auditor (A), Network Admin (NA) and Network Operator (NO), Security Admin (SA) and Security Operator (SO), LB Admin / LB Operator, VPN Admin, GI Partner Administrator and NETX Partner Administrator (service insertion partners), Support Bundle Collector, plus Cloud roles for cloud environments. Roles can be cloned and the clones edited; custom roles cannot reach features reserved to enterprise_admin or auditor.

For firewall operations assign Security Admin (configure DFW, Gateway Firewall, IDS/IPS, Malware Prevention objects) and Security Operator (read and operate, no configuration); Auditor is read-only across the system and suits compliance review. Network Admin cannot change security policy, which separates the networking and security personas the DFW categories are designed around (an application team in Application, network administrators in Infrastructure, security administrators in Emergency).

Projects (multi-tenancy): Project Admin has full access inside a project; Network Admin, Network Operator, Security Admin and Security Operator assigned in a project are scoped to that project only. Antrea groups are not available under projects.

Security Intelligence has its own RBAC: features are not exposed to roles without the corresponding permission; Enterprise Admin has full access (VCF 9.0.1 p.2376; Using and Managing Security Intelligence documentation).

Identity sources: local users, LDAP/Active Directory, and in VCF 9.0 the VCF Identity Broker / SSO integration; assign roles to groups rather than individuals and keep a local break-glass admin.

Key Takeaways

  • Security Admin configures, Security Operator operates, Auditor reads — Network Admin has no security rights
  • Custom roles are clones of built-ins and cannot gain enterprise_admin/auditor-only features
  • In projects the Admin/Operator roles are project-scoped; Project Admin is full access within the project

Exam Mapping: 6V0-21.25 — VMware vDefend Security for VCF 5.x Administrator

  • Official exam guide (Last Updated 05/12/2025): 17 weighted topics, summing to 100 %:
  • 01 Private Cloud Data Center Security (5 %)
  • 02 VMware vDefend Firewall Architecture (11 %)
  • 03 VMware vDefend Firewall Management (11 %)
  • 04 Lateral Protection with vDefend Distributed Firewall (7 %)
  • 05 Shared Services Platform (SSP) (2 %)
  • 06 Planning Application Segmentation with vDefend Security Intelligence (4 %)
  • 07 Context Aware Firewall and Identity Firewall (5 %)
  • 08 Protecting Container Workloads with vDefend Firewall (4 %)
  • 09 Gateway Firewall (7 %)
  • 10 Security Automation (5 %)
  • 11 Security Operations (2 %)
  • 12 Role-Based Access Control (4 %)
  • 13 Troubleshooting (4 %)
  • 14 Advanced Threat Prevention (2 %)
  • 15 IDPS (Intrusion Detection and Prevention System) (8 %)
  • 16 Malware Prevention Detection (8 %)
  • 17 NTA (Network Traffic Analysis) & NDR (Network Detection and Response) (11 %)
  • Topics 05, 06, 07, 08, 10 and 12 were not covered by the original theory and have been added to this page.
  • Site curriculum groups the guide into five study sections (Academy groupings with estimated weights, not the official blueprint): 1 vDefend Architecture (20 %), 2 Micro-Segmentation (30 %), 3 Threat Prevention (20 %), 4 Network Detection and Response (15 %), 5 Operations (15 %) — see the Curriculum tab for the objective list.

Labs in This Section

📝 Quiz — vDefend Security

0/55 correct

vDefend Architecture

Q1
Where does the vDefend Distributed Firewall enforce policy for each protected workload?
  • At the vNIC level inside the ESXi kernel (vfilter)
  • At the NSX Edge transport node
  • In a centralized service VM per cluster
  • On the physical ToR switch
The vDefend DFW enforces at the vNIC level inside the ESXi kernel using vfilter, inspecting traffic at the virtual NIC before it enters the guest. Not centralized per cluster, not on physical ToR, not on Edge nodes.
Q2
Which vDefend add-on is REQUIRED to enable Distributed IDS/IPS, Malware Prevention, and NDR analytics?
  • Firewall with ATP add-on
  • NSX Data Center Standard
  • Firewall add-on only
  • vSphere Foundation Advanced
The Firewall with ATP (Advanced Threat Prevention) add-on is required for Distributed IDS/IPS, Malware Prevention, and NDR analytics. Firewall add-on alone covers DFW only. NSX Standard and VVF Advanced lack ATP capabilities.
Q3
Which platform hosts scale-out advanced security services such as Malware Prevention and NDR?
  • Security Services Platform (SSP), formerly NSX Application Platform (NAPP)
  • SDDC Manager
  • VCF Operations Collector
  • vCenter Embedded PSC
The vDefend Security Services Platform (SSP) hosts the scale-out advanced security services; SSP 5.0 and later is the recommended platform, replacing the NSX Application Platform (NAPP) used in earlier NSX 4.x deployments (vDefend Firewall 4.2 documentation). SDDC Manager manages infrastructure. VCF Operations Collector gathers metrics. vCenter Embedded PSC handles identity.
Q4
Which statement best differentiates the Gateway Firewall (GFW) from the Distributed Firewall (DFW)?
  • GFW enforces on Tier-0/Tier-1 for north-south; DFW enforces at the vNIC for east-west
  • GFW only supports IPv6; DFW only supports IPv4
  • GFW is kernel based; DFW runs on Edge VMs
  • GFW uses IDS signatures; DFW does not
GFW enforces on Tier-0/Tier-1 gateways for north-south traffic. DFW enforces at the vNIC for east-west traffic. GFW is not kernel-based (it runs on Edge), and both support IDS when properly licensed.
Q5
Which component delivers the Central Control Plane configuration for DFW rules to ESXi transport nodes?
  • vCenter Server DRS
  • NSX Manager cluster via CCP/LCP
  • NSX Intelligence appliance
  • Edge transport node
NSX Manager cluster delivers the Central Control Plane (CCP/LCP) configuration for DFW rules to ESXi transport nodes. Edge nodes handle gateway services. vCenter DRS handles resource scheduling. NSX Intelligence provides analytics.

Micro-Segmentation

Threat Prevention

Network Detection and Response

Operations

🃏 Flashcards — vDefend Security

55 cards
Card 1 of 55
vDefend Platform
The VMware security product family layered on NSX that delivers DFW, Gateway Firewall, IDS/IPS, malware prevention, and NDR. It is licensed separately from NSX as Firewall or Firewall-with-ATP add-ons. It is the recommended east-west and workload-security stack in VCF 9.
Was this page useful?
Type to search. ↑ ↓ to move, Enter to open, Esc to close.