Group table

The group table introduces the ability to add support for port group abstraction for multi-pathing. This enables OpenFlow to represent a set of ports as a single entity for forwarding packets.

The group table supports the following group types:

  • All: Executes all the buckets in the group; mostly used for flooding and multicasting.
  • Indirect: Executes one defined bucket in the group. The action taken by this group type is sending packets to the next hop.
  • Select: Executes one bucket in the group. The action bucket is chosen by a switch-defined algorithm, such as round robin or hashing (for example, load sharing).
  • Fast failover: Executes the first live bucket, used in cases such as redundancy.

A group table consists of group entries. The counters in the following table are available in a group entry.

Group entry counters

Counter Description
Group Identifier A 32-bit unsigned integer uniquely identifying the group
Group type Determines group semantics
Counter Number of packets processed by a group
Action bucket

Ordered list of action buckets, where each action bucket contains a set of actions to execute and associated parameters

In this section:

  1. Scaling group numbers
    The output of the show openflow groups command displays the maximum number of actions in a bucket, the maximum number of buckets in a group and the maximum number of groups for scaling the group in OpenFlow.
  2. Considerations and limitations for group tables
    You must take into account the following when you configure group tables for OpenFlow flows.
  3. Group events
    The following group events are supported by OpenFlow.
Parent topic: OpenFlow v1.3.0