Energy-efficient network device with coordinated scheduling and processor rate control
View Patent ↗A network device of a communication network is configured to implement coordinated scheduling and processor rate control. In one aspect, packets are received in the network device and scheduled for processing from one or more queues of that device. An operating rate of a processor of the network device is controlled based at least in part on at least one of an arrival rate of the packets and a number of the packets stored in the one or more queues. As one example of processor rate control based on packet arrival rate, the operating rate of the processor may be controlled by accumulating a designated number of the received packets, determining an average arrival rate of the accumulated packets, and setting the operating rate of the processor based on the average arrival rate.
1. A method comprising the steps of:
receiving packets in a network device of a communication network;
scheduling the packets for processing from one or more queues of the network device; and
controlling an operating rate of a processor of the network device based at least in part on at least one of an arrival rate of the packets and a number of said packets stored in said one or more queues;
wherein the controlling step further comprises:
accumulating a designated number of the received packets;
determining an average arrival rate of the accumulated packets; and
setting the operating rate of the processor based on the average arrival rate; and
wherein the step of setting the operating rate of the processor based on the average arrival rate comprises setting the processing rate for an overlapped portion of two adjacent processing intervals to a combined operating rate for the two intervals subject to a maximum operating rate of the processor.
2. The method of claim 1 wherein the accumulating, determining and setting steps are repeated for each of a sequence of processing intervals.
3. The method of claim 2 wherein adjacent ones of the processing intervals are permitted to partially overlap with one another.
4. The method of claim 1 wherein the controlling step further comprises:
determining if at least a given one of the one or more queues is at or below a designated minimum fullness threshold;
if the given queue is at or below the designated minimum fullness threshold, setting the processor operating rate to a first relatively low level; and
subsequently increasing the processor operating rate from the first relatively low level to a second relatively high level.
5. The method of claim 4 wherein the first relatively low level is a minimum operating rate of the processor.
6. The method of claim 4 wherein the second relatively high level is the maximum operating rate of the processor.
7. The method of claim 4 wherein the step of increasing the processor operating rate comprising increasing the processor operating rate linearly as a function of time.
8. The method of claim 1 wherein the controlling step further comprises setting the operating rate of the processor as a designated function of the number of packets stored in the one or more queues.
9. The method of claim 8 wherein the designated function comprises a linear function.
10. The method of claim 1 wherein the step of scheduling the packets for processing further comprises scheduling the packets on a nearest-to-source basis such that one of the packets that has traveled a shorter distance from its source in the network is given priority over another one of the packets that has traveled a longer distance from its source in the network.
11. The method of claim 1 wherein the step of scheduling the packets for processing further comprises scheduling the packets on a farthest-to-go basis such that one of the packets that has a longer distance to travel to its destination in the network is given priority over another one of the packets that has a shorter distance to travel to its destination in the network.
12. The method of claim 1 wherein the step of scheduling the packets for processing further comprises scheduling the packets in accordance with a weighted fair queuing (WFQ) scheduling policy.
13. An article of manufacture comprising a non-transitory computer-readable storage medium having embodied therein executable program code that when executed by a processor of a network device causes the network device to perform the steps of:
receiving packets in the network device of a communication network;
scheduling the packets for processing from one or more queues of the network device; and
controlling an operating rate of the processor of the network device based at least in part on at least one of an arrival rate of the packets and a number of said packets stored in said one or more queues;
wherein the controlling step further comprises:
accumulating a designated number of the received packets;
determining an average arrival rate of the accumulated packets; and
setting the operating rate of the processor based on the average arrival rate; and
wherein the step of setting the operating rate of the processor based on the average arrival rate comprises setting the processing rate for an overlapped portion of two adjacent processing intervals to a combined operating rate for the two intervals subject to a maximum operating rate of the processor.
14. The article of manufacture of claim 13 wherein the controlling step further comprises:
determining if at least a given one of the one or more queues is at or below a designated minimum fullness threshold;
if the given queue is at or below the designated minimum fullness threshold, setting the processor operating rate to a first relatively low level; and
subsequently increasing the processor operating rate from the first relatively low level to a second relatively high level.
15. The article of manufacture of claim 13 wherein the controlling step further comprises setting the operating rate of the processor as a designated function of the number of packets stored in the one or more queues.
16. The article of manufacture of claim 13 wherein the step of scheduling the packets for processing further comprises one or more of:
scheduling the packets on a nearest-to-source basis such that one of the packets that has traveled a shorter distance from its source in the network is given priority over another one of the packets that has traveled a longer distance from its source in the network; and
scheduling the packets on a farthest-to-go basis such that one of the packets that has a longer distance to travel to its destination in the network is given priority over another one of the packets that has a shorter distance to travel to its destination in the network.
17. A network device comprising:
a processor having a controllable operating rate;
a memory coupled to the processor and comprising one or more queues;
network interface circuitry configured to support communication between the network device and one or more additional devices of a communication network; and
a control module;
wherein packets are received via the network interface circuitry and scheduled by the control module for processing from the one or more queues of the memory;
wherein the operating rate of the processor is controlled by the control module based at least in part on at least one of an arrival rate of the packets and a number of said packets stored in said one or more queues;
wherein the operating rate of the processor is controlled by:
accumulating a designated number of the received packets;
determining an average arrival rate of the accumulated packets; and
setting the operating rate of the processor based on the average arrival rate; and
wherein setting the operating rate of the processor based on the average arrival rate comprises setting the processing rate for an overlapped portion of two adjacent processing intervals to a combined operating rate for the two intervals subject to a maximum operating rate of the processor.
18. The network device of claim 17 wherein the operating rate of the processor is further controlled by determining if at least a given one of the one or more queues is at or below a designated minimum fullness threshold, and if the given queue is at or below the designated minimum fullness threshold, setting the processor operating rate to a first relatively low level, and subsequently increasing the processor operating rate from the first relatively low level to a second relatively high level.
19. The network device of claim 17 wherein the operating rate of the processor is further controlled by setting the operating rate of the processor as a designated function of the number of packets stored in the one or more queues.
20. The network device of claim 17 wherein the packets are scheduled for processing on one or more of:
a nearest-to-source basis such that one of the packets that has traveled a shorter distance from its source in the network is given priority over another one of the packets that has traveled a longer distance from its source in the network;
a farthest-to-go basis such that one of the packets that has a longer distance to travel to its destination in the network is given priority over another one of the packets that has a shorter distance to travel to its destination in the network; and
a weighted fair queuing basis.
21. The network device of claim 17 wherein the control module comprises a combined scheduler and rate adaptation module.
22. The network device of claim 21 wherein the control module is implemented at least in part in the form of software executed by the processor.