VMware ESXi and vSphere Cluster Management

Configure Traffic Shaping Policies on VMware ESXi Standard Switches

Learn how to configure outbound VM bandwidth limits on an ESXi standard vSwitch or port group using Average Bandwidth, Peak Bandwidth, and Burst Size.

ESXi standard-switch traffic shaping controls how quickly virtual machine network traffic is transmitted toward the physical network. This lesson explains the policy scope, bandwidth settings, burst behavior, configuration workflow, planning process, and troubleshooting steps.

What ESXi Traffic Shaping Does

A standard virtual switch, or standard vSwitch, is an ESXi host-local switch that connects virtual machines, VMkernel adapters, and physical network uplinks. A VM's virtual NIC (vNIC) connects to a port group, while a physical NIC (pNIC) provides an uplink toward the external network.

Without traffic shaping, a vNIC can transmit at the rate supported by available virtual-switch and uplink capacity. One busy workload can therefore consume capacity needed by other workloads. Traffic shaping adds a bandwidth-control policy that defines a sustained rate, a temporary peak rate, and a burst allowance.

Traffic shaping is disabled until an administrator explicitly enables and configures it. Common objectives include:

Preventing one workload from monopolizing uplink capacity.
Reducing noisy-neighbor behavior between VM workloads.
Allocating planned bandwidth to development, backup, replication, or application traffic.

Scope and Direction on a Standard vSwitch

The standard-vSwitch traffic shaping policy described here applies to outbound traffic: traffic transmitted from the ESXi virtual-switch context toward the physical network. It is not an inbound bandwidth limiter for traffic arriving at a VM.

You can configure the policy at either of two locations:

Standard vSwitch level: establishes a policy for the switch's attached port groups.
Port group level: establishes a more specific policy for workloads connected to that port group.

Policy inheritance means that a port group normally uses the setting inherited from its parent standard vSwitch. If the port group has its own traffic-shaping configuration, that configuration takes precedence over the inherited switch setting.

Configuration Location	Affected Objects	Precedence	Appropriate Use
Standard vSwitch	Attached port groups using the inherited policy	Parent default	Use when one baseline is suitable for all attached workloads.
Port group	VMs and other adapters connected to that port group	Overrides the vSwitch setting	Use for workload classes with distinct bandwidth requirements.

Traffic Shaping Parameters

The three settings use different units. Average Bandwidth and Peak Bandwidth are rates in kilobits per second (Kbps). Burst Size is a data allowance in kilobytes (KB). Bits and bytes are not interchangeable: eight bits equal one byte, so unit conversion is required when comparing these settings with application or physical-link measurements.

Setting	Unit	What It Controls	Configuration Rule
Average Bandwidth	Kbps	The sustained permitted transmission rate for a port.	Set it to the workload's normal long-term requirement.
Peak Bandwidth	Kbps	The maximum permitted transmission rate while a burst is available.	It must be greater than or equal to Average Bandwidth.
Burst Size	KB	The amount of data available for temporary burst transmission.	A larger value permits a larger short-lived burst, subject to the peak rate.

How the Values Work Together

Average Bandwidth establishes the long-term allowance.
Peak Bandwidth caps short-term transmission when the workload accelerates.
Burst Size determines how much data can use that temporary acceleration.

A workload may briefly need more throughput than its normal average, such as when an application returns a group of responses or a backup job sends a short block of data. Burst capacity allows that temporary increase, but it is not a permanent higher-bandwidth entitlement.

A small burst size limits how long the workload can accelerate. A larger burst size permits a larger short-term transfer, but the transfer remains subject to Peak Bandwidth, physical uplink capacity, and competition from other traffic. A burst setting cannot guarantee that the physical network will deliver the configured peak.

Choose the Policy Scope

Use a vSwitch-level policy when all attached port groups should share the same baseline objective. This is useful when the switch serves a consistent class of traffic.

Use a port group-level policy when workloads have different service requirements. For example, development traffic may need a modest limit, while an application port group may need a higher sustained rate and controlled short bursts.

Map port groups to recognizable workload classes such as development, backup, replication, or application traffic. Document every port group override because it supersedes the standard vSwitch setting and can make apparently similar workloads behave differently.

Configure Traffic Shaping in the vSphere Client

Open the target ESXi host in the vSphere client.
Go to the host's networking management area. The exact labels can vary by vSphere version.
Select the target standard vSwitch for a switch-wide policy, or select the relevant port group for a workload-specific override.
Open Edit Settings for the selected networking object.
Locate the Traffic Shaping section.
Enable traffic shaping.
Enter Average Bandwidth in Kbps.
Enter Peak Bandwidth in Kbps. Ensure that it is not lower than Average Bandwidth.
Enter Burst Size in KB.
Save the configuration.
Recheck the selected object and the effective port group setting to verify the intended scope and inheritance behavior.

For background on building the networking object, see Create Standard Switches. If VLAN-backed port groups are involved, review Configure VLANs.

Plan Values Before Applying the Policy

Confirm the speed and actual available capacity of the physical NIC and upstream network.
Identify whether the measurement is in bits or bytes. The ESXi bandwidth fields use Kbps, while Burst Size uses KB.
Set Average Bandwidth to the workload's normal sustained operating requirement rather than its occasional maximum.
Set Peak Bandwidth at or above Average Bandwidth and within a reasonable share of available uplink capacity.
Choose a Burst Size large enough for useful short spikes but not so large that many workloads can contend aggressively at once.
Test the policy with representative outbound traffic and monitor network performance after saving it.

Workload Type	Average Rate Goal	Peak Rate Goal	Burst Size Goal	Reasoning
Development or test workload	Modest normal-use rate	Controlled increase above average	Limited	Restricts noncritical workloads while allowing brief test activity.
General application workload	Aligned with normal application traffic	Moderately above the baseline	Enough for short response spikes	Supports normal bursts without unrestricted sustained throughput.
Short-burst workload	Lower long-term rate	Higher, but within uplink capacity	Selected for the expected transfer size	Allows brief acceleration while preserving the long-term allocation.

Example: Default Policy with an Exception

Suppose a standard vSwitch receives a general policy suitable for most port groups. A replication port group then receives its own policy because replication has a different bandwidth requirement. VMs on that port group use the replication policy, not the inherited vSwitch policy. Record this override in the network design and operational documentation.

Example: Invalid Rate Relationship

If Average Bandwidth is set to 10,000 Kbps and Peak Bandwidth is set to 8,000 Kbps, the configuration is invalid. A peak ceiling cannot be below the sustained allowance. Increase Peak Bandwidth to at least 10,000 Kbps, or reduce Average Bandwidth.

Validate Direction and Effective Behavior

Generate outbound traffic from a VM connected to the target port group and compare its measured throughput with the configured values. Observe both the sustained rate and short-lived spikes. Also check host and network performance counters during the test.

Then distinguish traffic arriving at the VM from traffic transmitted by it. High inbound traffic does not show that outbound shaping failed; the standard-vSwitch policy covered here controls the outbound direction. Use an appropriate upstream physical-network policy or another supported design when inbound control is required.

Troubleshooting

The VM Is Not Rate-Limited

Verify that traffic shaping is enabled, not merely populated with values.
Confirm that the VM's vNIC is connected to the intended port group.
Check whether the port group has an explicit setting that overrides the vSwitch policy.
Verify that the measured traffic is outbound from the shaped ESXi switching context.
Confirm that you are inspecting the correct ESXi host and standard vSwitch.

The Configuration Cannot Be Saved

Check that Peak Bandwidth is not lower than Average Bandwidth.
Confirm that all entries are valid positive numeric values.
Recheck units: bandwidth is Kbps and burst size is KB.
Verify that your account has permission to edit host networking settings. See Assign Permissions if access is the issue.

The Workload Does Not Reach the Configured Peak

Determine whether the burst allowance has been exhausted or the burst size is too small for the test pattern.
Check pNIC speed, physical uplink capacity, contention, and upstream network limits.
Confirm that the workload can generate enough outbound traffic to exercise the policy.
Check guest operating-system limits and other external network controls.

Incoming Traffic Remains High

Confirm that the expected result is inbound limiting. The standard-vSwitch shaping behavior covered here is outbound only. Select a control point that supports inbound policy, such as an upstream physical network policy or another supported network design.

Settings Differ Between Workloads

Compare the standard vSwitch and each relevant port group configuration.
Identify port groups with explicit overrides.
Document which workloads share each port group.
Confirm that the selected policy scope matches the intended design.

Operational Cautions

Physical network capacity and competing traffic still limit the result, regardless of configured burst or peak values.
Direction, policy placement, and vSwitch type all matter when evaluating observed throughput.
A standard vSwitch is host-local, so review the policy on every relevant host when the same design spans multiple hosts.
Distributed switches can have different management workflows and capabilities. Check the documentation for the exact vSphere and ESXi version when working with distributed-switch traffic management or automation interfaces. For architectural background, see Distributed Switches Explained.

Exam-Relevant Notes

Standard-vSwitch traffic shaping controls outbound traffic.
Traffic shaping must be enabled before its values take effect.
Average Bandwidth and Peak Bandwidth use Kbps; Burst Size uses KB.
Peak Bandwidth must be greater than or equal to Average Bandwidth.
A port group-specific policy takes precedence over the inherited standard vSwitch policy.
Traffic shaping controls transmission; it does not add physical bandwidth or replace upstream quality-of-service design.

For broader ESXi administration practice, continue with the VMware ESXi Online Course.