Technical data

Web Configuration

Corinex AV200 CableLAN Adapter

When several traffic flows are sharing the same network, there is sometimes the

need to establish several levels of priority to guarantee that bandwidth-sensitive

applications such as video or telephony continue to work smoothly even under net-

work congestion.

The traffic classifier is a packet inspector that is able to recognize several patterns

inside an Ethernet frame and assign a different priority to each of them. To ensure

that the classification is done to the right type of packet, there is a trigger mechanism

preceding the actual classification. The trigger mechanism is also based on pattern

recognition in a given location of the Ethernet packet. The next picture depicts the

packet classification mechanism.

There is one offset, one bitmask and pattern for the trigger condition. The trigger

condition is useful to make sure that the Ethernet frame contains, for example, an

IP frame. To check this condition, the offset would need to be set to 16 and the bit-

mask to 0xFFFF. If the resulting pattern is 0x0800, then the Ethernet frame contains

an IP packet and the classification can be done to a known field.

There is another offset and bitmask for the classification condition. The resulting

value is compared with a set of patterns. If the value matches a given pattern, the

packet will be classified with the specified priority. If the value does not match any of

the patterns, it will get a default priority.

There is a set of pre-defined criteria that allow classifying traffic based on the

802.1p

field of the Ethernet packet or the TOS field of the IP packet.