Cost efficient performance statistics gathering using logarithmic indexing
View Patent ↗A system for a distributed computing network for managing instrumentation information from a plurality of network-connected managed entities. One or more instrumentation processes are coupled to each of the network-connected managed entities and configured to gather performance metric values. A first classification process is responsive to information describing a type of input/output operation being performed and is configured to select one or more counters based upon the type of input/output operation being performed. A second classification process is coupled to receive a performance metric value from the one or more instrumentation processes. The second classification process is responsive to a computed logarithm of a measured value of at least one metric to select and increment a particular counter of the one or more counters selected by the first classification process.
1. A computer-readable storage medium tangibly embodying a program of instructions executable by a machine wherein said program of instruction comprises a plurality of program codes for management of a distributed computing network, said program of instruction comprising:
program code for one or more instrumentation processes having interfaces for generating performance metric values;
program code for a data structure having a plurality of logarithmically indexed counters; and
program code for processes that are responsive to the completion of an input/output operation and coupled to the instrumentation processes to receive the value of one of the performance metrics and increment a selected one of the indexed counters based upon the logarithm of the received value.
2. The program of instruction of claim 1 wherein the counters represent a plurality of categories wherein each category is associated with a range spanning from a starting point to an ending point and wherein the starting point and ending point define a logarithmic range.
3. The program of instruction of claim 1 wherein the counter is selected based upon a base-2 logarithm of the value of the performance metric.
4. The program of instruction of claim 1 wherein the program code for data structure comprises:
a first field holding a value indicating the maximum measured value for a particular performance metric that has occurred over a measurement period.
5. The program of instruction of claim 4 wherein the particular performance parameter is selected from the group consisting of: input/output operations per second (IOPs), throughput, and response time.
6. The program of instruction of claim 1 wherein the program code for the data structure implements a plurality of sets of logarithmically indexed counters where each set is associated with a particular type of loading and wherein the program code for the processes that select one of the indexed counters also select a particular one of the set of counters based upon the type of loading presented by the completed input/output operation.
7. The program of instruction software product of claim 6 wherein the type of loading information used to select a particular one of the sets of counters comprises information selected from the group consisting of: input/output operation size, whether the completed input output operation is random or sequential, and whether the completed input/output operation is a read or a write operation.
8. A management system for a distributed computing network comprising:
a management console having interfaces for accessing instrumentation information from a plurality of network-connected managed entities;
one or more instrumentation processes coupled to each of the network-connected managed entities and configured to gather performance metric values;
a first classification process responsive to a information describing a type of input/output operation being performed and configured to select one or more counters based upon the type of input/output operation being performed; and
a second classification process coupled to receive a performance metric value from the one or more instrumentation processes, wherein the second classification process is responsive to a computed logarithm of a measured value of at least one metric to select and increment a particular counter of the one or more counters selected by the first classification process.
9. The management system of claim 8 further comprising a data structure implementing the plurality of counters.
10. The management system of claim 9 wherein the counters represent a plurality of categories wherein each category is associated with a range spanning from a starting point to an ending point and wherein the starting point and ending point define a logarithmic range.
11. The management system of claim 8 wherein the first classification process is responsive to information selected from the group consisting of:
input/output operation size, whether the completed input output operation is random or sequential, and whether the completed input/output operation is a read or a write operation.
12. The management system of claim 8 wherein the second classification process performs a base-2 logarithm of the measured value of the at least one metric.
13. The management system of claim 8 wherein the at least one measured value used by the second classification process comprises response time of the input/output operation being performed.
14. The management system of claim 8 wherein the at least one measured value used by the second classification process comprises average queue depth of the input/output operation being performed.
15. The management system of claim 8 wherein the at least one measured value used by the second classification process comprises command receive to command issue time of the input/output operation being performed.
16. A computer implemented method for collecting performance statistics comprising:
Accessing, by said computer, instrumentation information from a plurality of network-connected managed entities;
in response to completing a particular input/output operation, identifying a set of counters comprising one or more counters based upon the type of operation that was completed; and
in response to completing the particular input/output operation, selecting a particular counter from the identified set of counters based upon a computed logarithm of a measured value of at least one metric; and
incrementing the selected counter.
17. The method of claim 16 wherein the act of selecting a particular counter from the identified set of counters based upon a computed logarithm uses a base-2 logarithmic computation.
18. A data structure implemented in a memory of a computing device, the data structure being configured for collecting performance statistics related to activity performed by the computing device, wherein the data structure comprises:
one or more arrays where each array corresponds to a particular type of activity of the computing device;
a plurality of elements within each array, wherein each element of the array corresponds to a logarithmically defined range for a selected characteristic of the activity being performed by the computing device;
a plurality of fields within each element of the array, wherein each field holds one or more values of a particular one of the performance statistics being gathered.
19. The data structure of claim 18 wherein at least one field within an element comprises a sub-array, wherein the sub-array has a plurality of elements and each of the plurality of elements in the sub-array corresponds to a logarithmically defined range for a selected characteristic of a performance statistic being gathered.