System and method for continuously measuring quality of service in a federated application environment

• US 6,003,079 A
• Filed: 02/27/1997
• Issued: 12/14/1999
• Est. Priority Date: 02/27/1997
• Status: Expired due to Term

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett Packard Company (HP Inc.)
• Inventors: Rolia, Jerome, Friedrich, Richard J.
• Primary Examiner(s): Maung, Zarni
• Assistant Examiner(s): Cardone, Jason D.
• Application Number:
  US08/807,073
• Time in Patent Office: 1,020 Days
• Field of Search: 395/200.49, 395/200.48, 395/200.54, 395/184.01, 702/186, 702/176, 709/219, 709/218, 709/224, 709/223, 709/226, 709/105, 709/106, 709/107
• US Class Current: 709/224
• CPC Class Codes: H04L 41/5009 Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF]
• CPC Class Codes: H04L 43/091 Measuring contribution of individual network components to actual service level

Abstract

A system and method for measuring quality-of-service in a federated application environment is described. One or more administrative domains are provided with each administrative domain being interconnected with at least one other administrative domain to form the federated application environment. One or more interconnected network nodes are situated within each administrative domain with each network node including at least one capsule interface within which a managed method is executed. Instrumentation is associated with each capsule interface of each network node with the instrumentation collecting performance data on the managed method being executed within the capsule interface. A count sensor determines a processing time with the instrumentation within the capsule interface for each managed method. A network sensor determines send and receive bandwidth demand information with the instrumentation within the capsule interface for each managed method. A demand sensor determines physical host resource demands as needed to satisfy invocations of the managed method with the instrumentation within the capsule interface for each managed method. An interval sensor determines response times with the instrumentation within the capsule interface for each managed method. A display displays the throughput metrics, bandwidth demand metrics, service demand metrics and response time metrics.

Claims

1. A method for continuously measuring quality-of-service for a federated application environment having a plurality of interconnected network nodes, each network node including at least one capsule interface within which a managed method and instrumentation are executed, the method comprising:

determining throughput metrics by measuring processing times with the instrumentation within the capsule interface for each managed method, wherein the managed methods executed in the nodes together form a distributed application, wherein one of the throughput metrics is incremented when a response is sent to a network transport subsystem to avoid an overlap of metrics, wherein one of the throughput metrics is incremented before a request for service such that the metric is not affected by delta latencies;

determining bandwidth demand metrics by capturing send and receive bandwidth demand information with the instrumentation within the capsule interface for each managed method;

determining service demand metrics by capturing physical host resource demands as needed to satisfy invocations of the managed method with the instrumentation within the capsule interface for each managed method; and

determining response time metrics by measuring response times with the instrumentation within the capsule interface for each managed method.

2. A method according to claim 1, wherein the step of determining the throughput metrics further comprises the steps of:

measuring method throughput information with a count sensor into an arrival metric object;

measuring correlating method throughput information with a count sensor using service information for managed methods executed by remote capsule interfaces into a completion metric object; and

measuring throughput information with an array of count sensors for requests of the managed method for service from remote capsule interface methods into an array of remote visit metric objects.

3. A method according to claim 2, wherein the step of determining the bandwidth demand metrics further comprises the steps of:

determining count information with a network sensor for measuring use of bandwidth demand for interactions with clients of the managed method into a client network metric object; and

determining count information with an array of network sensors on use of bandwidth demand for interactions of the managed method with managed methods which are executing on remote capsule interfaces into an array of remote network metric objects.

4. A method according to claim 2, wherein the step of determining the service demand metrics further comprises the steps of:

measuring host resource consumption with a demand sensor for the managed method into a service demand metric object; and

measuring host transport protocol host resource consumption by the managed method with an array of demand sensors for interactions with clients of the managed method and each remote capsule interface methods of the managed method into an array of transport demand metric objects.

5. A method according to claim 2, wherein the step of determining the response time metrics further comprises the steps of:

measuring response time with an array of interval sensors for the managed method at each remote capsule interface method of the managed method into an array of remote response metric objects;

measuring elapsed host transport protocol time of the managed method with an array of interval sensors for interactions by the managed method with clients of the managed method and each remote capsule interface method of the managed method into an array of transport response metric objects;

measuring network media time of the managed method with an array of interval sensors for interactions by the managed method with each remote capsule interface method of the managed method into an array of network response metric objects;

measuring queueing delay of clients at the managed method with an interval sensor into a queue delay metric object;

measuring time at host resources for the managed method with an interval sensor into a local response metric object; and

measuring residence time of clients accessing the managed method with an interval sensor into a method residence metric object.

6. A method for identifying quality-of-service problems in a federated application environment, the federated application environment comprising one or more interconnected network nodes, each being operable to execute a managed method, comprising the steps of:

determining a throughput rate for the managed method by measuring processing times, wherein all managed methods executed in the nodes together form a distributed application, wherein the throughput rate is incremented when a response is sent to a network transport subsystem to avoid processing overlap, wherein the throughput rate is incremented before a request for service such that the rate is not affected by delta latencies;

reporting that the managed method is processing too many arrivals if the throughput rate is qualitatively too high;

determining a residence time for clients accessing the managed method if the throughput rate is qualitatively acceptable; and

performing problem diagnosis if the residence time is qualitatively unacceptable.

7. A method according to claim 6, further comprising the steps of:

determining a transport response to transport demand ratio, the transport response comprising a measurement of elapsed host transport protocol times for interactions by the managed method with clients of the managed method and with each managed method of remote interface capsules, the transport demand comprising a measurement of host transport protocol resource consumption for interactions by the managed method with clients of the managed method and with each managed method of remote interface capsules; and

identifying a problem with a network upon which the managed method operates if the transport response to the transport demand ratio is qualitatively high.

8. A method according to claim 7, wherein the step of identifying a problem further comprising the steps of:

determining local response to service demand ratio, the local response comprising a measurement of time spent by the managed method using resources of a local host network node, the service demand comprising a measurement of resource consumption by the manage method of the local host network node;

comparing the transport response to transport demand ratio to the local response to service demand ratio; and

reporting a problem with the local host network node upon which the managed method operates if the local response to the service demand ratio is qualitatively less than the transport response to the transport demand ratio.

9. A method according to claim 6, further comprising the steps of:

determining a local response to service demand ratio, the local response comprising a measurement of time spent by the managed method using resources of a local host network node, the service demand comprising a measurement of resource consumption by the manage method of the local host network node; and

reporting a problem with an interface capsule of the managed method not receiving adequate service from resources of the local host network node if the local response to the service demand ration is qualitatively too high.

10. A method according to claim 6, further comprising the steps of:

determining a queueing delay of clients using the managed method; and

identifying a problem with the managed method if the queueing delay is qualitatively too high.

11. A method according to claim 10, where in the step of identifying a problem further comprises the steps of:

determining a local response to service demand ratio, the local response comprising a measurement of time spent by the managed method using resources of a local host network node, the service demand comprising a measurement of resource consumption by the manage method of the local host network node; and

reporting a problem with the managed method not receiving sufficient internal concurrency support from a capsule interface of the managed method.

12. A method according to claim 6, further comprising the steps of:

determining a remote response for the managed method for a further managed method at a remote capsule interface of the managed method; and

reporting a problem with one such further managed method if the remote response is qualitatively poor for the further managed method.

13. A method according to claim 6, further comprising the steps of:

determining a network response for a network media time of the managed method for interactions with a further managed method at a remote capsule interface of the managed method; and

reporting a problem with a network upon which one such further managed method is situated if the network response is qualitatively poor.

14. A quality-of-service measurement system for a federated application environment having a plurality of interconnected network nodes and managed methods, each operating on one of the nodes, the system comprising:

a plurality sets of instrumentation, each set associated with one of the nodes to collect performance data on a managed method executed within the associated node, wherein all the managed methods together form a distributed application, wherein the instrumentation sets further comprise a delta instrumentation fix to reduce the impact of delta latencies on execution of the instrumentation sets;

a plurality of quality-of-service objects executable on each of the nodes to measure quality-of-service of a managed method within its associated node based on a standardized set of quality-of-service metrics and the performance data collected by the instrumentation for the managed method, wherein the quality-of-service objects cause one of the metrics to be incremented when a response is sent to a network transport subsystem so as to avoid an overlap of metrics, wherein the quality-of-service objects also cause one of the metrics to be incremented before a request for service such that the metric is not affected by delta latencies.

15. The quality-of-service measurement system of claim 14, wherein the quality-of-service objects comprise

a count sensor object that determines a processing time of the managed method based on the standardized set of quality-of-service metrics and the collected data;

a network sensor object that determines send and receive bandwidth demand information of the managed method based on the standardized set of quality-of-service metrics and the collected data;

a demand sensor object that determines physical host resource demands as needed to satisfy invocations of the managed method based on the standardized set of quality-of-service metrics and the collected data;

an interval sensor object that determines response times of the managed method based on the standardized set of quality-of-service metrics and the collected data.

16. The quality-of-service measurement system of claim 15, wherein the count sensor object further comprise

an arrival metric object that measures method throughput information;

a completion metric object that measures correlating method throughput information with service information of remote managed methods of the managed method;

a remote visit metric object that measures throughput information for requests of the managed method for service from remote managed methods of the managed method.

17. The quality-of-service measurement system of claim 15, wherein the network sensor object further comprises

a client network metric object that determines count information regarding use of bandwidth demand for interaction with clients of the managed method;

a remote network metric object that determines count information regarding use of bandwidth demand for interaction with remote methods of the managed method.

18. The quality-of-service measurement system of claim 15, wherein the demand sensor object further comprises

a service demand metric object that measures host transport protocol host resource consumption of the managed method for interactions with clients of the managed method and remote methods of the managed method.

19. The quality-of-service measurement system of claim 15, wherein the interval sensor object further comprises

a remote response metric object that measures response time for the managed method at each remote method of the managed method;

a transport response metric object that measures elapsed host transport protocol time of the managed method for interactions by the managed method with clients of the managed method and with each remote method of the managed method;

a network response metric object that measures network media time of the managed method for interactions by the managed method with each remote method of the managed method;

a queue delay metric object that measures queuing delay of clients at the managed method;

a local response metric object that measures time at host resources for the managed method; and

a method residence metric object that measures residence time of clients accessing the managed method.