Configuring Auto-bandwidth Bundle TE++

Table 1. Feature History Table

Feature Name

Release Information

Feature Description

Automatic Bandwidth Bundle TE++ for Numbered Tunnels

Release 7.10.1

We have optimized network performance and enabled efficient utilization of resources for numbered tunnels based on real-time traffic by automatically adding or removing tunnels between two endpoints. This is made possible because this release introduces support for auto-bandwidth TE++ for numbered tunnels, expanding upon the previous support for only named tunnels, letting you define explicit paths and allocate the bandwidth to each tunnel.

The feature introduces these changes:

MPLS-TE tunnels are used to set up labeled connectivity and to provide dynamic bandwidth capacity between endpoints. The auto-bandwidth feature addresses the dynamic bandwidth capacity demands by dynamically resizing the MPLS-TE tunnels based on the measured traffic loads. However, many customers require multiple auto-bandwidth tunnels between endpoints for load balancing and redundancy. When the aggregate bandwidth demand increases between two endpoints, you can either configure auto-bandwidth feature to resize the tunnels or create new tunnels and load balance the overall demand over all the tunnels between two endpoints. Similarly, when the aggregate bandwidth demand decreases between two endpoints you can either configure the auto-bandwidth feature to decrease the sizes of the tunnel or delete the new tunnels and load balance the traffic over the remaining tunnels between the endpoints. The autobandwidth bundle TE++ feature is an extension of the auto-bandwidth feature and allows you to automatically increase or decrease the number of MPLS-TE tunnels to a destination based on real time traffic needs.

Tunnels that are automatically created as a response to the increasing bandwidth demands are called clones. The cloned tunnels inherit properties of the main configured tunnel. However, user configured load interval cannot be inherited. The original tunnel and its clones are collectively called a set. You can specify an upper limit and lower limit on the number of clones that can be created for the original tunnel.

Splitting is the process of cloning a new tunnel when there is a demand for bandwidth increase. When the size of any of the tunnels in the set crosses a configured split bandwidth, then splitting is initiated and clone tunnels are created.

The following figure explains creating clone tunnels when the split bandwidth is exceeded.

Merging is the process of removing a clone tunnel when the bandwidth demand decreases. If the bandwidth goes below the configured merge bandwidth in any one of the tunnels in the set, clone tunnels are removed.

The following figure explains removing clone tunnels to merge with the original tunnel when the bandwidth falls below the merge bandwidth.

There are multiple ways to equally load-share the aggregate bandwidth demand among the tunnels in the set. This means that an algorithm is needed to choose the pair which satisfies the aggregate bandwidth requirements. You can configure a nominal bandwidth to guide the algorithm to determine the average bandwidths of the tunnels. If nominal bandwidth is not configured, TE uses the average of split and merge bandwidth as nominal bandwidth.

Restrictions and Usage Guidelines

The following usage guidelines apply for the auto-bandwidth bundle TE++ feature.

  • This feature is only supported for the named tunnels and not supported on tunnel-te interfaces.

  • The range for the lower limit on the number of clones is 0 to 63 and the default value for the lower limit on the number of clones is 0.

  • The range for the upper limit on the number of clones is 1 to 63 and the default value for the upper limit on the number of clones is 63.

Configuration Example

This example shows how to configure the autobandwidth bundle TE++ feature for a named MPLS-TE traffic tunnel. You should configure the following values for this feature to work:

  • min-clones: Specifies the minimum number of clone tunnels that the original tunnel can create.

  • max-clones: Specifies the maximum number of clone tunnels that the original tunnel can create.

  • nominal-bandwidth: Specifies the average bandwidth for computing the number of tunnels to satisfy the overall demand.

  • split-bandwidth: Specifies the bandwidth value for splitting the original tunnel. If the tunnel bandwidth exceeds the configured split bandwidth, clone tunnels are created.

  • merge-bandwidth: Specifies the bandwidth for merging clones with the original tunnel. If the bandwidth goes below the configured merge bandwidth, clone tunnels are removed.

In this example, the lower limit on the number of clones is configured as two and the upper limit on the number of clones is configured as four. The bandwidth size for splitting and merging is configured as 200 and 100 kbps. 


RP/0/RP0/CPU0:router(config)# mpls traffic-eng 
RP/0/RP0/CPU0:router(config-mpls-te)# named-tunnels 
RP/0/RP0/CPU0:router(config-te-named-tunnels)# tunnel-te xyz
RP/0/RP0/CPU0:router(config-te-tun-name)# auto-bw
RP/0/RP0/CPU0:router(config-mpls-te-tun-autobw)# auto-capacity
RP/0/RP0/CPU0:router(config-te-tun-autocapacity)# min-clones 2
RP/0/RP0/CPU0:router(config-te-tun-autocapacity)# max-clones 4
RP/0/RP0/CPU0:router(config-te-tun-autocapacity)# nominal-bandwidth 150
RP/0/RP0/CPU0:router(config-te-tun-autocapacity)# split-bandwidth 200
RP/0/RP0/CPU0:router(config-te-tun-autocapacity)# merge-bandwidth 100