Consider the following excerpts from the Enterprise Voice Deployment Guide.
From page 58:
Configuring a failover route
Extending the previous example, if an administrator wants to define a failover route that may be used in case the Dallas-GW1 is brought down for maintenance, or is unavailable, the following example illustrates the required configuration change:
|
User Policy |
Phone Usages |
|
Default Calling Policy |
Local
GlobalPSTNHopoff |
|
Redmond Local Policy |
RedmondLocal |
|
Dallas Calling Policy |
DallasUsers
GlobalPSTNHopoff |
|
Route Name |
Number Pattern |
Phone Usages |
Gateway |
|
Redmond Local Route |
^\+1(425|206|253)(\d{7})$ |
Local
RedmondLocal |
Red-GW1
Red-GW2 |
|
Dallas Local Route |
^\+1(972|214|469)(\d{7})$ |
Local |
Dallas-GW1 |
|
Universal Route |
^\+?(\d*)$ |
GlobalPSTNHopoff |
Red-GW1
Red-GW2
Dallas-GW1 |
|
Dallas Users Route |
^\+?(\d*)$ |
DallasUsers |
Dallas-GW1 |
In the previous example, a phone usage of GlobalPSTNHopoff is added after the DallasUsers phone usage in the Dallas Calling Policy.
This enables calls with the Dallas Calling policy to use routes that are configured for the GlobalPSTNHopoff if a route for DallasUsers phone usage is unavailable.
From page 23-24:
The following are examples of failover logic related to gateway selection:
When there are multiple gateways that serve a particular route, a round-robin algorithm is used to distribute the calls across the multiple gateways.
Each gateway is configured with a maximum number of failed call attempts before traffic to the gateway is throttled. The default number of attempts is 10, but this value can be changed using a WMI (Windows® Management Interface) script. For a particular call, a given gateway may not be attempted more than once. If all gateways that serve a particular route are marked as unavailable, the server drops the call and notifies the client. You can also configure a gateway to be removed from the selection logic for some period of time. The unresponsive gateway is removed from the list of available gateways for increasingly longer periods of time, up to a maximum of 60 minutes, during which time the server repeatedly attempts to elicit a positive response. Upon receiving a positive response, the server returns the gateway to the list of available gateways.
Based on this information I have several questions about route selection:
In what situation is a route marked as unavailable? I understand that a particular gateway can be marked as unavailable based on recent failed call attempts. Does this mean that a route is marked as unavailable if all gateways in the route are unavailable?
Is there a way to administratively disable a route to temporarily re-route traffic out a different gateway?
Where is the script that is used to modify the maximum number of failed call attempts for gateways? Is it possible to configure the length of time the gateway will be removed from the selection logic when the failed call threshold is reached?
Thanks in advance,
Jamie Schwinn
www.systmsny.net
