Self-adaptive scheduling method and network element
The present invention relates to a method and network element for scheduling data transmission in a packet data network, e.g. a differentiated services network. Data traffic of a predetermined traffic class is classified based on a predetermined traffic profile, e.g. responsive or non-responsive traffic, and an in-profile rate of the data traffic is estimated. Then, a weight determining a transmit order of queued data packets of said predetermined traffic class is adjusted according to the estimated in-profile rate. Thus, the data traffic is marked into different classes based on the traffic profile and a queue scheduler improves fairness between the classified flows in the packet data network. Thereby, enhanced fairness between traffic profiles can be achieved easily and effectively.
1. A method, comprising:
marking a data packet by setting a differentiated services code point value in a data packet header of the data packet;
classifying the data packet of an assured forwarding class based on a predetermined profile of different transport layer packets;
estimating, by periodically calculating by a processor, an in-profile data packet arrival rate including a green packet rate of said data packet at a predetermined interval, the calculating determined as at least one of a moving average or a sliding window;
queuing said data packet in a separate sub-queue of a plurality of sub-queues based on said in-profile data packet arrival rate;
adjusting a weight according to said in-profile data packet arrival rate of said queued data packet, said weight determining a transmit order of said queued data packet;
allocating bandwidth to each of the sub-queues proportional to said adjusted weight; and
scheduling a data transmission of a plurality of predetermined traffic classes based on said adjusted weight.
2. The method according to claim 1 , further comprising using a fair scheduling procedure in said scheduling step, said fair scheduling procedure being configured to allocate bandwidth to each sub-queue proportional to said adjusted weight.
3. The method according to claim 2 , wherein said fair scheduling procedure is a weighted round robin, weighted fair queuing, virtual clock, or self-clocked fair queuing procedure.
4. The method according to claim 1 , wherein said predetermined data packet profile is used to separate responsive traffic from non-responsive traffic.
5. The method according to claim 4 , wherein said responsive traffic is transmit control protocol traffic and said non-responsive traffic is user datagram protocol traffic.
6. The method according to claim 1 , wherein said in-profile rate is a green packet rate of three color marker marked packets.
7. The method according to claim 1 , wherein said adjusting is performed periodically at a predetermined or dynamic interval.
8. The method according to claim 1 , further comprising using a priority scheduling for an expedited forwarding class traffic, and sharing remaining bandwidth by said scheduling method.
9. An apparatus, comprising:
a marker configured to mark a data packet by setting a differentiated services code point value in a data packet header of the data packet;
a classifier configured to classify the data packet of an assured forwarding class based on a predetermined profile of different transport layer packets;
an estimator configured to estimate, by periodically calculating, an in-profile data packet arrival rate including a green packet rate of said data packet at a predetermined interval, the calculating determined as at least one of a moving average or a sliding window; and
a scheduler configured to adjust a weight according to said in-profile data packet arrival rate, wherein bandwidth is allocated to each of the sub-queues based on the adjusted weight;
wherein said classifier is further configured to queue said data packet in a separate sub-queue based on said in-profile data packet arrival rate;
wherein said weight determines a transmit order of said queued data packet; and
wherein said scheduler is further configured to schedule a data transmission of a plurality of predetermined traffic classes based on said adjusted weight.
10. The apparatus of claim 9 , wherein said apparatus is a core router of a differentiated service network.
11. A non-transitory computer readable storage medium encoded with instructions that, when executed by a computer, perform:
marking a data packet by setting a differentiated services code point value in a data packet header of the data packet;
classifying the data packet of an assured forwarding class based on a predetermined profile of different transport layer packets;
estimating, by periodically calculating, an in-profile data packet arrival rate including a green packet rate of said data packet at a predetermined interval, the calculating determined as at least one of a moving average or a sliding window;
queuing said data packet in a separate sub-queue of a plurality of sub-queues based on said in-profile data packet arrival rate;
adjusting a weight according to said in-profile data packet arrival rate of said queued data packet, said weight determining a transmit order of said queued data packet;
allocating bandwidth to each of the sub-queues proportional to said adjusted weight; and
scheduling a data transmission of a plurality of predetermined traffic classes based on said adjusted weight.
12. The non-transitory computer readable storage medium according to claim 11 , wherein the instructions, when executed by the computer, further perform:
using a fair scheduling procedure in said scheduling step, said fair scheduling procedure being configured to allocate bandwidth to each sub-queue proportional to said adjusted weight.
13. The non-transitory computer readable storage medium according to claim 12 , wherein said fair scheduling procedure is a weighted round robin, weighted fair queuing, virtual clock, or self-clocked fair queuing procedure.
14. The non-transitory computer readable storage medium according to claim 11 , wherein said predetermined data packet profile is used to separate responsive traffic from non-responsive traffic.
15. The non-transitory computer readable storage medium according to claim 14 , wherein said responsive traffic is transmit control protocol traffic and said non-responsive traffic is user datagram protocol traffic.
16. The non-transitory computer readable storage medium according to claim 11 , wherein said in-profile rate is a green packet rate of three color marker marked packets.
17. The non-transitory computer readable storage medium according to claim 11 , wherein said adjusting is performed periodically at a predetermined or dynamic interval.
18. The non-transitory computer readable storage medium according to claim 11 , wherein the instructions, when executed by the computer, further perform:
using a priority scheduling for an expedited forwarding class traffic, and sharing remaining bandwidth by said scheduling method.