Advanced features through API extensions¶

VMware NSX QoS extension¶

The VMware NSX QoS extension rate-limits network ports to guarantee a specific amount of bandwidth for each port. This extension, by default, is only accessible by a project with an admin role but is configurable through the policy.json file. To use this extension, create a queue and specify the min/max bandwidth rates (kbps) and optionally set the QoS Marking and DSCP value (if your network fabric uses these values to make forwarding decisions). Once created, you can associate a queue with a network. Then, when ports are created on that network they are automatically created and associated with the specific queue size that was associated with the network. Because one size queue for a every port on a network might not be optimal, a scaling factor from the nova flavor rxtx_factor is passed in from Compute when creating the port to scale the queue.

Lastly, if you want to set a specific baseline QoS policy for the amount of bandwidth a single port can use (unless a network queue is specified with the network a port is created on) a default queue can be created in Networking which then causes ports created to be associated with a queue of that size times the rxtx scaling factor. Note that after a network or default queue is specified, queues are added to ports that are subsequently created but are not added to existing ports.

$ neutron queue-create --min 10 --max 1000 myqueue

$ neutron net-create network --queue_id QUEUE_ID

$ neutron queue-create --default True --min 10 --max 2000 default

$ neutron queue-list

$ neutron queue-delete QUEUE_ID_OR_NAME

VMware NSX provider networks extension¶

Provider networks can be implemented in different ways by the underlying NSX platform.

The FLAT and VLAN network types use bridged transport connectors. These network types enable the attachment of large number of ports. To handle the increased scale, the NSX plug-in can back a single OpenStack Network with a chain of NSX logical switches. You can specify the maximum number of ports on each logical switch in this chain on the max_lp_per_bridged_ls parameter, which has a default value of 5,000.

The recommended value for this parameter varies with the NSX version running in the back-end, as shown in the following table.

Recommended values for max_lp_per_bridged_ls

In addition to these network types, the NSX plug-in also supports a special l3_ext network type, which maps external networks to specific NSX gateway services as discussed in the next section.

VMware NSX L3 extension¶

NSX exposes its L3 capabilities through gateway services which are usually configured out of band from OpenStack. To use NSX with L3 capabilities, first create an L3 gateway service in the NSX Manager. Next, in /etc/neutron/plugins/vmware/nsx.ini set default_l3_gw_service_uuid to this value. By default, routers are mapped to this gateway service.

$ openstack network create public --external --provider-network-type l3_ext \
--provider-physical-network L3_GATEWAY_SERVICE_UUID

$ openstack network create public --external --provider-network-type l3_ext \
--provider-physical-network L3_GATEWAY_SERVICE_UUID --provider-segment VLAN_ID

Operational status synchronization in the VMware NSX plug-in¶

Starting with the Havana release, the VMware NSX plug-in provides an asynchronous mechanism for retrieving the operational status for neutron resources from the NSX back-end; this applies to network, port, and router resources.

The back-end is polled periodically and the status for every resource is retrieved; then the status in the Networking database is updated only for the resources for which a status change occurred. As operational status is now retrieved asynchronously, performance for GET operations is consistently improved.

Data to retrieve from the back-end are divided in chunks in order to avoid expensive API requests; this is achieved leveraging NSX APIs response paging capabilities. The minimum chunk size can be specified using a configuration option; the actual chunk size is then determined dynamically according to: total number of resources to retrieve, interval between two synchronization task runs, minimum delay between two subsequent requests to the NSX back-end.

The operational status synchronization can be tuned or disabled using the configuration options reported in this table; it is however worth noting that the default values work fine in most cases.

When running multiple OpenStack Networking server instances, the status synchronization task should not run on every node; doing so sends unnecessary traffic to the NSX back-end and performs unnecessary DB operations. Set the state_sync_interval configuration option to a non-zero value exclusively on a node designated for back-end status synchronization.

The fields=status parameter in Networking API requests always triggers an explicit query to the NSX back end, even when you enable asynchronous state synchronization. For example, GET /v2.0/networks/NET_ID?fields=status&fields=name.

Big Switch router rules¶

Big Switch allows router rules to be added to each project router. These rules can be used to enforce routing policies such as denying traffic between subnets or traffic to external networks. By enforcing these at the router level, network segmentation policies can be enforced across many VMs that have differing security groups.

$ neutron router-update ROUTER_UUID --router_rules type=dict list=true \
  source=any,destination=any,action=permit \
  source=external,destination=10.10.10.0/24,action=deny

$ neutron router-update ROUTER_UUID --router_rules type=dict list=true  \
  source=any,destination=any,action=permit \
  source=10.10.10.0/24,destination=any,action=permit,nexthops=10.10.10.254+10.10.10.253

$ neutron router-update ROUTER_UUID --router_rules type=dict list=true \
  source=any,destination=any,action=permit \
  source=10.10.10.0/24,destination=10.20.20.20/24,action=deny