Understanding Fleet Latency in VMware Cloud Foundation 9.0

Network latency plays a critical role in designing and operating a modern multi-instance cloud environment. With the introduction of enhanced multi-instance capabilities in VMware Cloud Foundation (VCF) 9.0, understanding fleet latency is now more important than ever for architects and administrators.

To support this, VMware has published a VCF Fleet Latency Network Diagram (available on the VMware Ports & Protocols site), which provides clear guidance on latency thresholds across core components.

What is Fleet Latency?

Fleet latency refers to the time taken for communication between:

• VCF instances

• Management domains

• Workload domains

• Core platform components

In a distributed VCF deployment, latency directly impacts:

• Lifecycle management operations

• Monitoring and telemetry

• Automation workflows

• Cross-instance orchestration

While VMware defines maximum supported latency thresholds, these values represent upper limits—not ideal targets. Designing for lower latency always results in better performance and user experience.

Control Plane vs Data Plane: Why Latency Impacts Differ

Not all traffic behaves the same under latency. Understanding this distinction is key:

Control Plane Traffic (UI/API)

• Typically asynchronous

• Requests are received quickly but responses may be delayed

• Impact: Slower UI operations or API responses, but workflows still complete

Data Plane Traffic (Binary Transfers)

• Includes downloads/uploads (e.g., patch bundles)

• Highly sensitive to latency

• Impact: Reduced throughput and longer transfer times

Practical Insight: For lifecycle operations, binaries can be staged in advance. This reduces the real-time impact of latency during upgrades or patching activities.

Network Architecture Overview

A typical VCF Fleet setup consists of multiple instances. Each instance includes:

Management Domain Components

• vCenter Server – centralized infrastructure management

• SDDC Manager – lifecycle and configuration control

• NSX Manager & Edge – network virtualization and routing

Workload Domains

• Dedicated vCenter instances

• ESXi hosts running application workloads

  
  

Platform Layers

Automation Layer

• Handles orchestration across instances

• Sensitive to latency for workflow execution

Operations Layer

• Includes telemetry collectors and analytics engines

• Responsible for monitoring, alerting, and insights

Key Latency Considerations

The VCF Fleet Latency Diagram highlights communication paths between components. These paths influence:

• Automation workflows (orchestration across instances)

• Monitoring and telemetry collection

• Cross-instance operations

Important Threshold Example

One critical requirement to note:

• VCF Operations Collector → Core Components (vCenter, NSX, SDDC Manager): ≤ 50 ms

Real-World Scenario: Brownfield Import

If a workload domain (e.g., imported environment) is geographically distant:

• Latency may exceed 50 ms

• This can impact telemetry collection

Recommended Approach:

• Deploy an additional local Operations Collector

• Keep ≤50 ms locally

• Allow up to ≤300 ms back to the central Operations layer

This design ensures both compliance and performance.

Best Practices for Designing Low-Latency VCF Fleets

1. Optimize Network Design

• Use high-bandwidth, low-latency links between sites

• Prefer direct connectivity over complex routing paths

2. Segment Traffic

Separate traffic types for better efficiency:

• Management

• Workload

• Operations

3. Stage Lifecycle Content

• Pre-download upgrade bundles

• Avoid real-time dependency on high-latency links

  
  

4. Monitor Continuously

• Track latency between components

• Identify bottlenecks early

  
  

5. Build for Resilience

• Implement redundancy and failover

• Ensure connectivity during outages

The introduction of fleet-level architecture in VMware Cloud Foundation 9.0 brings powerful capabilities but also new design considerations. Latency is no longer just a network metric; it directly influences automation, operations, and lifecycle success.

The newly published latency diagram is a strong starting point, covering core VCF components. As the platform evolves, additional components and scenarios are expected to be included.

If you’re designing or operating a VCF fleet, now is the time to revisit your network architecture and ensure it aligns with these latency guidelines.