Product Documentation

Priority Queuing

Jun 14, 2016

The priority queuing feature lets you filter incoming HTTP traffic on the basis of categories that you create and define, and prioritize those HTTP requests accordingly. Priority queuing directs high-priority requests to the server ahead of low-priority requests, so that users who need resources for important business uses receive expedited access to your protected Web servers.

Note: The priority queuing feature is not supported in NetScaler 9.2 nCore.

To implement priority queuing, you create priority queuing policies that specify a priority, weight, threshold, and implicit action. When an incoming request matches a priority queuing policy, the request is processed as the associated action indicates. For example, you can create a priority queuing policy that places all matching requests above a certain threshold in a surge queue, while giving priority treatment to other requests.

You can bind up to three priority queuing policies to a single load balancing virtual server. The priority levels are:

Level 1
A Level 1 policy processes priority requests.
Level 2
A Level 2 policy processes requests that should receive responses as soon as Level 1 requests have been cleared from the queue.
Level 3
A Level 3 policy processes non-priority requests that receive responses only after requests in the first two queues have been cleared.

You can use weighted queuing to adjust the relative priority of each of these queues. Weights can range from 0 to 101. A weight of 101 tells the NetScaler appliance to clear all requests in that queue before forwarding any requests in the lower-priority queues to the Web server. A weight of 0 tells the appliance to send requests in that queue to the Web server only when there are no requests waiting in any of the other queues.

You must assign a unique name to each priority queuing policy. Policy names can be up to 127 characters. Multiple policies bound to the same load balancing virtual server cannot have the same priority level. No two virtual servers that have one or more common underlying physical services can have priority queuing configured or enabled on both virtual servers simultaneously.

To configure priority queuing the NetScaler, you perform the following steps:

  • Enable the load balancing feature
  • Define a server and service
  • Define a load balancing virtual server
  • Bind the service to the load balancing virtual server
  • Enable the priority queuing feature
  • Create the priority queuing policies
  • Bind the priority queuing policies to the load balancing virtual server
  • Enable priority queuing on load balancing virtual server

For information about enabling load balancing, creating servers, creating virtual servers and services, and binding these servers and services, see the Citrix NetScaler Traffic Management Guide at http://support.citrix.com/article/CTX132359. For complete information about policies and expressions, see the Citrix NetScaler Policy Configuration and Reference Guide at http://support.citrix.com/article/CTX132362.