| 摘要 |
The present invention relates to a scheduling unit adapted to aggregate data units from a plurality of input ports to a common output port according to a service sequence. The service sequence identifies which particular input port must be served for a particular service slot. A scheduling unit according to the present invention features optimized jitter and queue occupancy figures. This is achieved by allotting a number of service slots to each input port in proportion to its service share, next by determining for each input port a corresponding number of candidate positions uniformly spaced over the service sequence, finally by resolving any contention between the so-determined candidate positions. A service slot is assigned first and foremost to the input port which has the highest number of candidate positions considered while determining the service sequence yet not assigned to any service slot, next which gets the highest service share.
|
