Systems and methods for reducing power consumption by monitoring network traffic
A method can include monitoring a level of network traffic to a server. The method can include setting a clock rate of one or more cores of a processor of the server based on the monitored level of network traffic.
1. A computer-implemented method comprising:
monitoring a level of network traffic to a server including measuring a ratio of user space execution to kernel space execution of a network driver of the server by:
measuring a first ratio of user space execution to kernel space execution of a first network driver of a first core of the one or more cores; and
measuring a second ratio of user space execution to kernel space execution of a second network driver of a second core of the one or more cores; and
setting a clock rate of one or more cores of a processor of the server based on the monitored level of network traffic,
wherein setting the clock rate of the one or more cores based on the monitored level of network traffic comprises setting the clock rate of the first core based on the first ratio and setting the clock rate of the second core based on the second ratio.
2. The computer-implemented method of claim 1 , wherein setting the clock rate of the one or more cores based on the level of network traffic comprises reducing the clock rate based on the level of network traffic being below a predetermined threshold.
3. The computer-implemented method of claim 2 , wherein reducing the clock rate includes maintaining a reserve of available processing power on the one or more cores.
4. The computer-implemented method of claim 2 , wherein setting the clock rate of the one or more cores based on the level of network traffic further comprises increasing the clock rate based on the level of network traffic being above the predetermined threshold.
5. The computer-implemented method of claim 2 , further comprising increasing the clock rate based on a buffer of a network driver of the server being filled above a predetermined buffer threshold, wherein the buffer provides capacity for short-term traffic variations and indicates when traffic exceeds processing capacity.
6. The computer-implemented method of claim 1 , wherein setting the clock rate of the one or more cores comprises reducing the clock rate of a first core of the one or more cores, the method further comprising increasing the clock rate of the first core based on demand for a specific function performed by the first core.
7. The computer-implemented method of claim 6 , wherein the specific function performed by the first core includes at least one of encryption, decryption, compression, decompression, tunneling encapsulation, tunneling de-encapsulation, packet storage to disk, and network firewall.
8. A computer-implemented method comprising:
monitoring a level of network traffic to a server including measuring a ratio of user space execution to kernel space execution of a network driver of the server by:
measuring a first ratio of user space execution to kernel space execution of a first network driver of a first core of the one or more cores; and
measuring a second ratio of user space execution to kernel space execution of a second network driver of a second core of the one or more cores; and
setting a clock rate of one or more cores of a processor of the server based on the monitored level of network traffic,
wherein setting the clock rate of the one or more cores based on the monitored level of network traffic comprises setting the clock rate of the first core and the second core based on a maximum of the first ratio and the second ratio.
9. A computer-implemented method comprising:
generating, using a predictive model, a predicted level of network traffic to a server, wherein the predictive model is a machine learning model trained using historical data including historical levels of network traffic to the server and a ratio of user space execution to kernel space execution of a network driver of the server; and
setting a clock rate of one or more cores of a processor of the server based on the predicted level of network traffic,
wherein the historical levels of network traffic include core-level data including a first ratio of user space execution to kernel space execution of a first network driver of a first core of the one or more cores and a second ratio of user space execution to kernel space execution of a second network driver of a second core of the one or more cores, and wherein the historical data on which the machine learning model is trained includes the core-level data.
10. The computer-implemented method of claim 9 , wherein reducing the clock rate includes maintaining a reserve of available processing power on the one or more cores.
11. The computer-implemented method of claim 9 , wherein setting the clock rate of the one or more cores comprises increasing the clock rate based on a buffer of a network driver of the server being filled above a predetermined buffer threshold, wherein the buffer provides capacity for short-term traffic variations and indicates when traffic exceeds processing capacity.
12. The computer-implemented method of claim 9 , wherein setting the clock rate of the one or more cores comprises reducing the clock rate of a first core of the one or more cores, the method further comprising increasing the clock rate of the first core based on demand for a specific function performed by the first core.
13. The computer-implemented method of claim 12 , wherein the specific function performed by the first core includes at least one of encryption, decryption, compression, decompression, tunneling encapsulation, tunneling de-encapsulation, packet storage to disk, and network firewall.
14. The computer-implemented method of claim 9 , wherein the historical data on which the machine learning model is trained includes processor-level data comprising a maximum of the first ratio and the second ratio.
15. The computer-implemented method of claim 1 , wherein the historical data includes at least one trend fitted to the levels of network traffic to the server.