Method and apparatus for using multiple network processors to achieve higher performance networking applications

A method, comprising:

• a) classifying a plurality of packets with a plurality of network processors, each of said network processors classifying a different packet of said plurality of packets by examining look-up information stored in one of a plurality of look-up resources coupled thereto, with the look-up information contained in each of said plurality of look-up resources being the same and with the look-up resource coupled to one of the plurality of network processors differing from the look-up resource coupled to the remaining network processors of the plurality of network processors, the classifying including identification of a queue from amongst a plurality of queues for queuing each packet;
• b) storing different pieces of a first packet of said plurality packets into different memory units of a plurality of memory units, each of said different memory units being managed by corresponding memory controllers, wherein each of said different memory units corresponds to a different dimension of each queue amongst the plurality of queues;
• and
• c) storing different pieces of a second packet of said plurality of packets into different memory units of said plurality of memory units, wherein a last piece of the first packet and a first piece of the second packet share a same memory unit after said storings, resulting in different memory controllers managing the first piece of the second packet and a remainder of the second packet.

An apparatus, comprising:

• a plurality of memory units managed by corresponding memory controllers;
• a plurality of queues;
• and
• a plurality of network processors that classify a plurality of packets, each of said network processors classifying a different packet of said plurality of packets by examining look-up information stored in one of a plurality of look-up resources coupled thereto, with the look-up information contained in each of said plurality of look-up resources being the same and with the look-up resource coupled to one of the plurality of network processors differing from the look-up resource coupled to the remaining network processors of the plurality of network processors, the classification including identification of a queue from the plurality of queues for queuing the different packet;
• wherein different pieces of the different packet are stored into different memory units, each of said different memory units corresponding to a different dimension of each queue amongst the plurality of queues;
• wherein header information units are distributed to the plurality of network processors prior to said classification, each header information unit comprising header information from one of said plurality of packets.

A method, comprising:

• a) distributing header information units to a plurality of network processors, each header information unit comprising header information from one of a plurality of packets;
• b) classifying the plurality of packets with the plurality of network processors, each of said network processors classifying a different packet of said plurality of packets by examining look-up information stored in one of a plurality of look-up resources coupled thereto, with the look-up information contained in each of said plurality of look-up resources being the same and with the look-up resource coupled to one of the plurality of network processors differing from the look-up resource coupled to the remaining network processors of the plurality of network processors, the classifying including identification of a queue from amongst a plurality of queues for queuing the different packet;
• and
• c) storing different pieces of one of said packets into different memory units, each of said different memory units being coupled to a different one of said network processors, wherein a last piece of a first packet and a first piece of a second packet share a same memory unit after being stored resulting in different memory controllers managing the first piece of the second packet and a remainder of the second packet, wherein each of said different memory units corresponds to a different dimension of each queue amongst the plurality of queues.