Method and apparatus for resource allocation

A method for packet processing in a network device, the method comprising:

• queuing packets in a plurality of queues that share a processing resource for packet processing, the plurality of queues including a first queue;
• generating tokens respectively associated with the queues when the queues change from empty to non-empty;
• storing tokens in a buffer and outputting the tokens from the buffer in a first-in-first-out manner;
• storing respective queue status information within respective tokens, each queue status information regarding respective status of the respective queue associated with the respective token;
• in response to outputting a first token associated with the first queue, outputting from the first queue a first packet for processing by the processing resource;
• and
• in response to determining that first queue status information in the first token that was output from the buffer indicates that the first queue stores one or more additional packets, re-store the first token to permit the first queue to output another packet in connection with the first token again being output from the buffer.

A resource arbiter for scheduling allocation of a shared resource for serving a plurality of client devices, the resource arbiter comprising:

• a first-in-first-out (FIFO) selection circuit configured to:
• receive tokens respectively associated with the plurality of client devices which request the shared resource, a token associated with a client device being indicative of an allocation of the shared resource to the client device when the token is output by the FIFO selection circuit, and
• store the tokens in a buffer and output the tokens from the buffer in a first-in-first-out manner,
• store respective queue status information within respective tokens, each queue status information regarding respective status of the respective queue associated with the respective token, and
• in response to determining that first queue status information in a first token output from the buffer indicates that a first client device requires an additional resource allocation of the shared resource in excess of an allocation of the shared resource to the first client device as indicated by outputting of the first token from the buffer, re-store the first token in the buffer to permit the first client device to receive the additional resource allocation of the shared resource in connection with the first token again being output from the buffer.