...
Info |
---|
This document describes how to set a bandwidth limit in a virtual datacenter for each |
...
public IP using |
...
quality of |
...
service (QoS) traffic shaping |
...
parameters. |
...
|
This feature applies to public IPs in infrastructure and managed by NSX in vCenter and vCenter clusters.
Panel | ||
---|---|---|
| ||
Privilege: Manage bandwidth limit for public IPs |
During an upgrade or when you create a virtual datacenter, the new "publicbandwidthlimit" attribute is present but set to 0. You must public IP bandwidth limit is disabled.
To use QoS, edit the limit to set a value and . The platform will apply the value when you deploy or reconfigure a VM to use a public IP in a deployed VM to enable the limit in the virtual datacenter and apply it in the cloud..
...
To edit the bandwidth limit:
Select the virtual datacenter
...
Go to Network → QoS
Click the pencil edit button
To
...
activate the bandwidth limit in a specific direction, select the Enabled checkbox for that direction
...
For the Average, enter the amount of bandwidth, in bits per second, that each public IP in the virtual datacenter can use
For the Peak, enter the maximum bandwidth in bits per second that each public IP in the virtual datacenter can use
For the Burst size, enter the amount of data that can be transmitted at the peak bandwidth rate in bytes. A burst bonus accumulates when traffic is below the Average value and this bandwidth can be used for bursts
...
...
To register changes that were made outside the platform, save existing NAT public IP bandwidth values.
In the API, to register changes, send a POST request with the existing values.
For more details about using NAT in Abiquo, see:
.
This feature sets the limits when you deploy or reconfigure a VM to use a public IP.
If you deploy a VM and the platform cannot configure the limit, then the deploy will fail and roll back. If you reconfigure a VM and try to add a public IP, and the limit fails, then the platform will roll back the reconfigure and delete the public IP address.
If you are working with multiple VMs and there are different switches involved, then the platform will make a best effort to update all of the VMs, and log any errors.