Packet Matching

This section describes how packet matching occurs at the SDN switch.

Packet match fields are extracted from the packet when it arrives at the switch. These fields—which can include packet header information such as destination IP address—or the ingress port are used to do a flow table lookup. A match is found if the values in the packet match fields used for the lookup match those defined in the flow table entry. A match can also result if a flow table entry has a value of ANY, which causes the entry to match all values in the packet header. You can use a bitmask on match fields to match a broader set of packets.

When more than one match is found for a packet, the highest priority match is selected, the counters associated with the selected flow entry are updated, and the instruction set from the selected flow entry is applied.

Note: If multiple matching flow entries with the same highest priority are found, the selection method of the flow entry is undefined. This can happen if a controller writer never sets the CHECK_OVERLAP bit on flow modification messages, resulting in overlapping entries.

If no matches are found in the flow table entries, the packet is matched to the table-miss entry of the table (see Table-Miss Flow Entry).

Figure 473: Packet Matching Workflow

Packet Matching Workflow

For more information, refer to the following sections: