6V0-22.25 Planning and Designing

Planning and Designing Detailed Explanation

Capacity Planning and Service Engine Sizing

Exam Radar

• Core Priority: Accurately assessing throughput and transaction requirements to determine physical resource allocation.

• High Frequency: Accounting for the CPU overhead of SSL/TLS handshakes, comparing RSA and EC certificate impacts.

• Confusion Alert: Distinguishing between per-SE resource limits (vCPU/Memory) and Service Engine Group (SEG) aggregate limitations.

• Scenario Logic: Calculating required SE resources based on expected Gigabits per second (Gbps) and Connections Per Second (CPS).

• Version Delta: Modern sizing must account for the specific performance profile of the underlying vSphere host hardware.

Atomic Deconstruction

• Actionable: Conduct a sizing exercise by evaluating the application's peak throughput and cryptographic requirements before defining SE Group resource maximums.

• Parametric: Monitor the impact of Elliptic Curve (EC) certificates, which typically offer lower CPU overhead compared to traditional RSA certificates at similar security levels.

• Causal: Under-provisioning vCPU or Memory at the SE level will lead to packet drops and increased latency when the application reaches peak CPS.

SKILLS.md Matrix

Metric	Design Constraint	Impact
Throughput	Gbps requirement	Number of SEs / vCPU count
Transactions	CPS requirement	CPU cycles for handshakes
Memory	Concurrent sessions	Persistence table and buffer size

Elastic Scale-Out and Scale-In Logic

Exam Radar

• Core Priority: Defining the automated triggers that adjust data plane capacity without manual intervention.

• High Frequency: Setting CPU, memory, and throughput thresholds that signal the Controller to provision additional SEs.

• Confusion Alert: Distinguishing between the scale-out event (adding SEs) and the cooling period (preventing rapid flapping).

• Scenario Logic: Predicting how the Controller redistributes existing connections and manages Gratuitous ARP during a scale-out.

• Version Delta: Utilizing duration-based rules to ensure stability during fluctuating traffic patterns.

Atomic Deconstruction

• Actionable: Configure Analytics Profiles with specific percentage thresholds (e.g., 80 percent CPU) to trigger the automated spin-up of new SE VMs.

• Parametric: Define scale-in cooling periods to ensure that capacity is not removed too quickly after a transient traffic spike.

• Causal: Automatic scale-out prevents application downtime by ensuring that the Virtual Service (VS) always has sufficient aggregate CPU/RAM to process incoming requests.

SKILLS.md Matrix

Logic	Design Requirement	Goal
Scale-Out	High-watermark threshold	Performance preservation
Scale-In	Cooling period duration	Flapping prevention
Redundancy	N+M Buffer	Spare capacity maintenance

High Availability (HA) Mode Selection

Exam Radar

• Core Priority: Aligning the HA mode with application criticality and Recovery Time Objectives (RTO).

• High Frequency: Comparing Elastic HA (N+M) for efficient redundancy versus Elastic HA (Active/Active) for maximum throughput.

• Confusion Alert: Identifying that Legacy Active/Standby is suitable only for basic requirements and limits total throughput.

• Scenario Logic: Choosing the Active/Active design for instantaneous failover where multiple SEs serve the same VIP simultaneously.

• Version Delta: Selecting the default recommended N+M mode for a balance of efficiency and availability.

Atomic Deconstruction

• Actionable: Select the HA mode within the SE Group configuration based on whether the application requires maximum throughput (Active/Active) or efficient resource usage (N+M).

• Parametric: In N+M mode, configure the "M" buffer to specify how many spare SEs the Controller should maintain across the group.

• Causal: Active/Active HA ensures that if a single SE fails, the remaining SEs continue to process traffic for the Virtual Service with zero transition time.

SKILLS.md Matrix

HA Mode	Capability	Use Case
Legacy	Active/Standby	Simple, low-throughput apps
Elastic N+M	Buffer-based recovery	Standard enterprise apps
Active/Active	Simultaneous processing	Mission-critical, high-scale apps

Virtual Service and Pool Design Best Practices

Exam Radar

• Core Priority: Proper association of health monitors and persistence profiles to ensure predictable traffic flow.

• High Frequency: Choosing appropriate monitor types (HTTP, TCP, or Custom) to verify application layer responsiveness.

• Confusion Alert: Evaluating the need for HTTP Cookie persistence versus Source IP persistence in multi-server environments.

• Scenario Logic: Designing port usage standards such as port 80 to 443 redirection and selecting optimized TCP/UDP profiles.

• Version Delta: Emphasis on verifying backend responsiveness rather than just port connectivity.

Atomic Deconstruction

• Actionable: Select and configure health monitors that validate specific application strings or status codes to ensure pool members are truly functional.

• Parametric: Implement persistence profiles based on application session requirements to maintain client affinity to specific backend servers.

• Causal: Using HTTP Cookie persistence ensures that clients remain tied to the same backend server even if their source IP changes, which is critical for stateful applications.

SKILLS.md Matrix

Element	Atomic Requirement	Operational Detail
Health Monitoring	Response Validation	Beyond simple L4 port checks
Persistence	Session Continuity	Cookie vs. Source IP profiles
Traffic Flow	Redirection Logic	Standardized 80 to 443 mapping

Analytics Profiles and Service Engine Resource Impact

Exam Radar

• Core Priority: Balancing deep visibility requirements against Service Engine (SE) CPU and memory consumption.

• High Frequency: Designing log policies that default to Significant logs and use Full logs only for active troubleshooting.

• Confusion Alert: Distinguishing between the frequency of metric collection and the real-time impact of Client Insights on SE throughput.

• Scenario Logic: Adjusting the frequency of metric collection to reduce processing overhead on the SE data plane.

• Version Delta: High-fidelity logging requires a calculated design trade-off in resource-constrained environments.

Atomic Deconstruction

• Actionable: Configure Analytics Profiles to collect Significant logs by default to preserve SE resources while maintaining essential visibility.

• Parametric: Adjust metric collection intervals, moving from 5-second to 30-second intervals for non-critical services to lower CPU overhead.

• Causal: Reducing the log fidelity for stable virtual services frees up SE processing cycles for actual traffic handling and SSL termination.

SKILLS.md Matrix

Feature	Logic	Atomic Detail
Log Levels	Significant vs. Full	Resource-conscious telemetry
Metrics	Collection Frequency	Adjustable processing overhead
Client Insights	Real-User Monitoring	Impact on SE throughput capacity

GSLB and Multi-Site Architectural Planning

Exam Radar

• Core Priority: Coordinated design between global and local load balancing layers for multi-site availability.

• High Frequency: Architecting the Controller cluster to manage global configuration (Leader) while maintaining local site autonomy (Follower).

• Scenario Logic: Utilizing Geolocation databases to route clients to the nearest active site and minimize latency.

• Version Delta: Implementing split-brain prevention through health monitor logic that ensures safe failure states.

Atomic Deconstruction

• Actionable: Designate specific Controller clusters as GSLB Leaders to synchronize global application states across geographically dispersed sites.

• Parametric: Configure proximity-based steering within GSLB pools to automatically direct users based on their geographic location.

• Causal: Proper site health monitoring ensures that if a data center fails, DNS responses are automatically updated to point to the next closest healthy site.

SKILLS.md Matrix

Component	Logic	Detail
Site Roles	Leader/Follower	Configuration vs. Local Execution
Steering	Proximity-Based	Geolocation database integration
Resiliency	Fail-Safe Monitors	Split-brain and failover protection

GSLB Site Management and Leader/Follower Architecture

Exam Radar

• Core Priority: Defining the hierarchy of Controller clusters to manage global application delivery.

• High Frequency: Understanding the synchronization of global configuration from the GSLB Leader to all Follower sites.

• Confusion Alert: Distinguishing between a site's local load balancing autonomy and its participation in the global health fabric.

• Scenario Logic: Maintaining global availability when the GSLB Leader is unreachable by relying on local Follower site persistence.

• Version Delta: Strategic placement of the GSLB Leader in the most stable management domain to ensure global configuration integrity.

Atomic Deconstruction

• Actionable: Designate one Avi Controller cluster as the GSLB Leader to serve as the authoritative source for global virtual service configurations.

• Parametric: Configure Follower sites to receive and execute global policies while maintaining local control over their specific Service Engine Groups.

• Causal: Using a Leader/Follower architecture ensures that a localized failure in one data center does not disrupt the global DNS resolution or traffic steering logic for other healthy sites.

SKILLS.md Matrix

Role	Responsibility	Data Flow
GSLB Leader	Global Config Authoring	Push to all Follower sites
GSLB Follower	Local Policy Execution	Pull from Leader/Local Health Checks
Health Monitor	Cross-site Verification	Bi-directional site status sync

DNS Steering and Proximity-Based Planning

Exam Radar

• Core Priority: Selecting the optimal traffic steering algorithm based on user location and site health.

• High Frequency: Utilizing Geolocation databases to minimize latency by directing clients to the closest active data center.

• Scenario Logic: Implementing health-based steering where DNS records are automatically updated to exclude failed sites.

• Version Delta: Integration with internal or external DNS providers to automate the assignment of A/AAAA records for global VIPs.

Atomic Deconstruction

• Actionable: Integrate a Geolocation database into the Avi GSLB configuration to enable proximity-based steering for global applications.

• Parametric: Define "Active" and "Passive" site weights within the GSLB pool to control traffic distribution during normal operations and failover.

• Causal: DNS-based steering allows for global load balancing without requiring complex BGP Anycast configurations, simplifying the multi-site network architecture.

SKILLS.md Matrix

Steering Method	Operational Requirement	Primary Benefit
Proximity	Geolocation Database	Lowest client latency
Weighted	Traffic Ratio Config	Controlled canary/blue-green deploys
Failover	Priority-based Pools	Disaster recovery automation

Split-Brain Prevention and Fail-Safe Design

Exam Radar

• Core Priority: Designing monitor logic that prevents multiple sites from claiming "Active" status during a network partition.

• High Frequency: Implementing "fail-safe" or "fail-to-last-known-good" settings for DNS responses during site isolation.

• Scenario Logic: Predicting site behavior when synchronization between the GSLB Leader and Followers is lost.

• Version Delta: Modern health monitoring includes verifying the integrity of the data path, not just the management plane.

Atomic Deconstruction

• Actionable: Configure cross-site health monitors that verify connectivity between all GSLB sites to detect network partitions.

• Parametric: Set the GSLB DNS response policy to "return local" or a specific "failover IP" if the site becomes isolated from the Leader.

• Causal: Implementing split-brain prevention ensures that users are not directed to a site that is operational but isolated from the application database or backend services.

SKILLS.md Matrix

Condition	Logic	System Response
Site Isolation	Fail-safe policy	Maintain local vs. stop DNS
Leader Failure	Follower Autonomy	Local VS continues to function
Partition	Split-brain logic	Site health consensus required