Length-of-the-curve stress metric for improved characterization of computer system reliability
View Patent ↗Embodiments of the present invention provide a system that characterizes the reliability of a computer system. The system first collects samples of a performance parameter from the computer system. Next, the system computes the length of a line between the samples, wherein the line includes a component which is proportionate to a difference between values of the samples and a component which is proportionate to a time interval between the samples. The system then adds the computed length to a cumulative length variable which can be used to characterize the reliability of the computer system.
1. A method for characterizing the reliability of a computer system, comprising:
collecting samples of a performance parameter from the computer system;
computing a length of a line between the samples, wherein the line includes a component which is proportionate to a difference between values of the samples and a component which is proportionate to a time interval between the samples; and
adding the computed length to a cumulative length variable which can be used to characterize the reliability of the computer system.
2. The method of claim 1 , wherein computing the length of the line between samples involves computing √{square root over (|S 1− S 2 | 2 +t 2 )}, where S 1 and S 2 are the magnitudes of the samples and t is the magnitude of the time intervals between the samples.
3. The method of claim 1 , wherein collecting the samples involves collecting the samples at a predetermined time interval.
4. The method of claim 1 , wherein the performance parameter is a physical parameter, which includes at least one of: a temperature; a relative humidity; a cumulative or differential vibration; a fan speed; an acoustic signal; a current; a voltage; a time-domain reflectometry (TDR) reading; or another physical property that indicates an aspect of performance of the system.
5. The method of claim 1 , wherein the performance parameter is a software metric, which includes at least one of: a system throughput; a transaction latency; a queue length; a load on a central processing unit; a load on a memory; a load on a cache; I/O traffic; a bus saturation metric; FIFO overflow statistics; or another software metric that indicates an aspect of performance of the system.
6. The method of claim 1 , wherein the method further comprises analyzing the cumulative length variable to determine a likelihood of a failure in the computer system.
7. The method of claim 1 , wherein the method further comprises adjusting the computed length according to a function of the magnitude of the samples of the computer system performance parameter.
8. The method of claim 7 , wherein adjusting the computed length involves multiplying the computed length by a weight factor, wherein the weight factor is W=e 0.1(S 1.01 −373) +1, where S 1 and S 2 are the magnitudes of the samples and S=(S 1 +S 2 )/2.
9. An apparatus for characterizing the reliability of a computer system, comprising:
a collection mechanism configured to collect samples of a performance parameter from the computer system;
a computation mechanism configured to compute a length of a line between the samples, wherein the line includes a component which is proportionate to a difference between values of the samples and a component which is proportionate to a time interval between the samples; and
an adding mechanism configured to add the computed length to a cumulative length variable which can be used to characterize the reliability of the computer system.
10. The apparatus of claim 9 , wherein when computing the length of the line between samples, the computation mechanism is configured to compute √{square root over (|S 1 −S 2 | 2 +t 2 )}, where S 1 and S 2 are the magnitudes of the samples and t is the magnitude of the time intervals between the samples.
11. The apparatus of claim 9 , wherein when collecting the samples, the collection mechanism is configured to collect the samples at a predetermined time interval.
12. The apparatus of claim 9 , wherein the performance parameter is a physical parameter, which includes at least one of: a temperature; a relative humidity; a cumulative or differential vibration; a fan speed; an acoustic signal; a current; a voltage; a time-domain reflectometry (TDR) reading; or another physical property that indicates an aspect of performance of the system.
13. The apparatus of claim 9 , wherein the performance parameter is a software metric, which includes at least one of: a system throughput; a transaction latency; a queue length; a load on a central processing unit; a load on a memory; a load on a cache; I/O traffic; a bus saturation metric; FIFO overflow statistics; or another software metric that indicates an aspect of performance of the system.
14. The apparatus of claim 9 , further comprising an analyzing mechanism configured to analyze the cumulative length variable to determine a likelihood of a failure in the computer system.
15. The apparatus of claim 9 , wherein the computation mechanism is configured to adjust the computed length according to a function of the magnitude of the samples of the computer system performance parameter.
16. The apparatus of claim 15 , wherein when adjusting the computed length, the computation mechanism is configured to multiply the computed length by a weight factor, wherein the weight factor is W=e 0.1(S 1.01 −373) +1, where S 1 and S 2 are the magnitudes of the samples and S=(S 1 +S 2 )/2.
17. A computer system that characterizes reliability, comprising:
a processor;
a memory coupled to the processor, wherein the memory is configured to store instructions and data for the processor;
a collection mechanism configured to collect samples of a performance parameter from the computer system;
a computation mechanism configured to compute a length of a line between the samples, wherein the line includes a component which is proportionate to a difference between values of the samples and a component which is proportionate to a time interval between the samples; and
an adding mechanism configured to add the computed length to a cumulative length variable which can be used to characterize the reliability of the computer system.
18. The computer system of claim 17 , wherein when computing the length between samples, the computation mechanism is configured to compute √{square root over (|S 1 −S 2 | 2 +t 2 )}, where S 1 and S 2 are the magnitudes of the samples and t is the magnitude of the time intervals between the samples.
19. The computer system of claim 17 , wherein the computation mechanism is configured to adjust the computed length according to a function of the magnitude of the samples of the computer system performance parameter.
20. The computer system of claim 19 , wherein when adjusting the computed length, the computation mechanism is configured to multiply the computed length by a weight factor, wherein the weight factor is W=e 0.1(S 1.01 −373) +1, where S 1 and S 2 are the magnitudes of the samples and S=(S 1 +S 2 )/2.