Predictive model-based measurement acquisition employing a predictive model operating on a manager system and a managed system

• US 6,430,615 B1
• Filed: 03/13/1998
• Issued: 08/06/2002
• Est. Priority Date: 03/13/1998
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Hellerstein, Joseph L., Haus, Nikolaus
• Primary Examiner(s): Najjar, Saleh
• Application Number:
  US09/042,060
• Time in Patent Office: 1,607 Days
• Field of Search: 709/202, 709/217, 709/224, 709/226, 709/235
• US Class Current: 709/224
• CPC Class Codes: H04L 41/0233 Object-oriented techniques, for representation of network management data, e.g. common object request broker architecture [CORBA]
• CPC Class Codes: H04L 41/046 comprising network management agents or mobile agents therefor
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 43/00 Arrangements for monitoring or testing data switching networks
• CPC Class Codes: H04L 43/022 by sampling
• CPC Class Codes: H04L 43/106 using time related information in packets, e.g. by adding timestamps
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/101 based on network conditions
• CPC Class Codes: H04L 67/1023 based on a hash applied to IP addresses or costs
• CPC Class Codes: H04L 67/1029 using data related to the state of servers by a load balancer

Abstract

A method, system and program storage device for measurement acquisition using predictive models which: (a) can improve scalability; (b) can accommodate devices that operate in a disconnected mode; and (c) enable integration of data from sources with different time granularities. Various features can be embodied in software and an object-oriented implementation is described. Different implementations are described, such as standalone predictive models implemented only on a manager (for example systems management/load balancing) or managed system (for example router management); or a parallel implementation with predictive models running on both the manager and managed (agent) systems (for example financial trading or system health monitoring). In a parallel model implementation, the agent constructs a predictive model that is conveyed to the manager system. The models are used in parallel, possibly with no communication for an extended time. The manager uses its model to provide tentative values of measurement variables to management applications. The agent uses its model to check its accuracy. If the model is found to be insufficiently accurate, an updated model is transmitted to the manager. The invention allows other measurement acquisition protocols to operate concurrently on the same measurement variables.

Claims

1. A predictive model-based method for measurement data acquisition in a distributed system, comprising the steps of:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system;

operating the predictive model on both the manager system and the managed system; and

the manager system processing requests for the measurement variable values based on the predictive model on the manager system.

2. The method of claim 1, further comprising the steps of:

the disconnectable managed system determining that the predictive model is insufficiently accurate; and

the disconnectable managed system constructing and communicating an updated predictive model to the managed system, in response to said determining.

3. The method of claim 2, further comprising the step of, the managed system disconnecting from the managed system except to update the predictive model.

4. The method of claim 1, further comprising the step of communicating a measurement variable subscription request for a predictive model from the manager system to the managed system, wherein the predictive model is constructed in response to the subscription request.

5. The method of claim 1, further comprising the step of:

the managed system disconnecting from the managed system;

the managed system measuring one or more actual values of the measurement variables and comparing the actual value against an accuracy bound;

if the actual values are within an accuracy bound associated with the predictive model on the managed system;

the managed system connecting and communicating a message to the manager system confirming that the actual values are within the accuracy bound; and

if the actual values are not within the accuracy bound associated with the predictive model on the managed system;

the managed system constructing an updated predictive model; and

connecting and communicating the updated predictive model to the manager system.

6. The method of claim 1, further comprising the steps of:

the managed system measuring one or more actual values of the measurement variables and comparing an actual value against an accuracy bound; and

confirming and communicating a message to the manager system indicating the one or more actual values are within an accuracy bound.

7. The method of claim 6, further comprising the steps of:

the manager system associating said measurement variable values with one or more status codes indicating how they were obtained; and

the manager system storing said measurement variable values and the status codes, in response to said associating.

8. The method of claim 7, further comprising the steps of:

the manager system identifying a measurement variable value status as tentative if the measurement variable value is a predicted value which has not been confirmed as within the accuracy bound; and

the manager system changing the measurement variable value status from tentative to confirmed, in response to the message indicating the one or more actual values are within the accuracy bound.

9. The method of claim 7, further comprising the steps of:

the managed system measuring one or more actual values of said measurement variables;

the managed system communicating the actual values to the manager system; and

the manager system identifying the measurement variable value status as actual if the actual values have been received.

10. The method of claim 1, wherein the system includes multiple managed systems that have one or more of different collection frequencies and unsynchronized clocks, said method further comprising the steps of:

said dynamically constructing including creating and associating a time stamp with a measurement variable value;

the manager system synchronizing data from the multiple managed systems, using the time stamp.

11. The method of claim 10, wherein said step of the manager system synchronizing data further comprises the step of adjusting a granularity of the time stamp using the predictive models.

12. The method of claim 10, wherein the unsynchronized clocks operate at a same rate, further comprising the step of:

the manager system predicting measurement variable values at sampling intervals based on a relative time stamp indexing a set of sampling intervals.

13. The method of claim 1, further comprising the step of:

the managed system and the manager system synchronizing the measurement variable values used as inputs to the predictive model.

14. The method of claim 1, further comprising the step of:

manager system applications requesting and using the predicted values of measurement variables without requiring actual values of these variables from the managed system.

15. The method of claim 1, in an object-oriented system, further comprising the steps of:

said dynamically constructing comprises serializing and communicating serialized predictive model objects on the managed system and reconstructing a same predictive model on the manager system based on the serialized predictive model objects.

16. The method of claim 1, further comprising the step of transparently running multiple measurement acquisition protocols concurrently in the system.

17. The method of claim 1, further comprising the step of:

constructing and storing the predictive model as an expression tree.

18. The method of claim 1, further comprising the steps of:

the disconnectable managed system monitoring measured values and using the predictive model on the disconnectable managed system to detect excessive deviations of predicted values from measured values;

the disconnectable managed system communicating to the manager system updates to the predictive model if said excessive deviations are detected; and

the manager system updating its predictive model based on the updates.

19. The method of claim 1, further comprising the step of:

said dynamically constructing further comprises the step of dynamically creating the predictive model wherein the model includes a model definition, model parameters and model measurement variable inputs.

20. The method of claim 19, in an object-oriented system, said dynamically constructing further comprising the steps of:

the managed system communicating the predictive model to the manager system;

the manager system parsing the model definition, the model parameters, and the model inputs, in response to said communicating; and

the manager system creating and initializing objects for the model definition, the model parameters, and the model inputs.

21. A predictive model-based method for measurement data acquisition in a distributed system, comprising the steps of:

a manager application specifying measurement variables for which a subscription is requested;

a manager system communicating the subscription request and a desired accuracy of measurement variables to a managed system, in response to said specifying;

the manager system and the managed system storing measured values from the managed system, in response to the subscription;

the managed system creating and communicating a predictive model to the manager system, in response to said storing;

the manager system satisfying requests for measurement variable values using tentative predictive values from the predictive model on the manager system; and

the disconnectable managed system;

periodically checking the accuracy of the tentative predictive values;

communicating to the manager messages confirming the tentative predictive values that are within the desired accuracy;

updating the predictive model if the tentative predictive values are not within the desired accuracy and communicating an updated predictive model to the manager system.

22. The method of claim 21, further comprising the step of:

the manager system receiving one or more actual measurement variable values from the measurement data access facilities on the managed system; and

identifying and storing the actual values on the manager system.

23. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for a predictive model-based method for measurement data acquisition in a distributed system, said method steps comprising:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system;

operating the predictive model on both the manager system and the managed system; and

the manager system processing requests for the measurement variable values based on the predictive model on the manager system.

24. The program storage device of claim 23, further comprising the steps of:

the disconnectable managed system determining that the predictive model is insufficiently accurate; and

the disconnectable managed system constructing and communicating an updated predictive model to the managed system, in response to said determining.

25. The program storage device of claim 23, further comprising the step of communicating a measurement variable subscription request for a predictive model from the manager system to the managed system, wherein the predictive model is constructed in response to the subscription request.

26. The program storage device of claim 23, further comprising the step of:

the managed system measuring one or more actual values of the measurement variables and comparing an actual value against an accuracy bound; and

confirming and communicating a message to the manager system indicating the one or more actual values are within an accuracy bound.

27. The program storage device of claim 26, further comprising the steps of:

the manager system associating said measurement variable values with one or more status codes indicating how they were obtained; and

the manager system storing said measurement variable values and the status codes, in response to said associating.

28. The program storage device of claim 27, further comprising the steps of:

the manager system identifying a measurement variable value status as tentative if the measurement variable value is a predicted value which has not been confirmed as within the accuracy bound; and

the manager system changing the measurement variable value status from tentative to confirmed, in response to the message indicating the one or more actual values are within the accuracy bound.

29. The program storage device of claim 23, wherein the system includes multiple managed systems that have one or more of different collection frequencies and unsynchronized clocks, said method further comprising the steps of:

said dynamically constructing including creating and associating a time stamp with a measurement variable value; and

the manager system synchronizing data from the multiple managed systems, using the time stamp.

30. The program storage device of claim 29, wherein said step of the manager system synchronizing data further comprises the step of adjusting a granularity of the time stamp using the predictive models.

31. The program storage device of claim 29, wherein the unsynchronized clocks operate at a same rate, further comprising the step of:

the manager system predicting measurement variable values at sampling intervals based on a relative time stamp indexing a set of sampling intervals.

32. The program storage device of claim 23, in an object-oriented system, further comprising the steps of:

said dynamically constructing comprises serializing and communicating serialized predictive model objects on the managed system and reconstructing a same predictive model on the manager system based on the serialized predictive model objects.

33. The program storage device of claim 23, further comprising the step of transparently running multiple measurement acquisition protocols concurrently in the system.

34. The program storage device of claim 23, further comprising the step of:

said dynamically constructing further comprises the step of dynamically creating the predictive model wherein the model includes a model definition, model parameters and model measurement variable inputs.

35. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for predictive model-based method for measurement data acquisition in a distributed system, said method steps comprising:

a manager application specifying measurement variables for which a subscription is requested;

a manager system communicating the subscription request and a desired accuracy of measurement variables to a managed system, in response to said specifying;

the manager system and the managed system storing measured values from the managed system, in response to the subscription;

the managed system creating and communicating a predictive model to the manager system, in response to said storing;

the manager system satisfying requests for measurement variable values using tentative predictive values from the predictive model on the manager system; and

the disconnectable managed system;

periodically checking the accuracy of the tentative predictive values;

communicating to the manager messages confirming the tentative predictive values that are within the desired accuracy;

updating the predictive model if the tentative predictive values are not within the desired accuracy and communicating an updated predictive model to the manager system.

36. The program storage device of claim 35, further comprising the step of:

the manager system receiving one or more actual measurement variable values from the measurement data access facilities on the managed system; and

identifying and storing the actual values on the manager system.

37. A predictive model-based measurement acquisition system for creating and managing one or more predictive models that predict values of measurement variables in distributed systems, said system comprising:

a manager system;

a disconnectable managed system;

means for dynamically constructing and coordinating operation of a predictive model on both the disconnectable managed system and the manager system, the predictive model adapted for predicting values of said measurement variables; and

manager system means for processing requests for the measurement variable values based on the predictive model on the manager system.

38. The system of claim 37, further comprising:

the disconnectable managed system including;

agent memory means for storing the predictive model;

an agent data access facility means for providing measured values of said measurement variables; and

agent measurement repository means, coupled to the agent data access facility, for storing measurement variable values known to the manager; and

an agent model handler, coupled to the agent measurement repository, for updating and operating the predictive model on the disconnectable managed system;

the manager system including;

manager memory means for storing the predictive model;

manager measurement repository means for storing the measurement variable values in the memory;

one or more management applications, coupled to the memory, for requesting and using predicted values generated on the manager system and processing measurement variable values; and

manager model handler means, coupled to the memory, for operating the predictive model on the manager and updating the predictive model based on information communicated from the managed system.

39. A predictive model-based method for measurement data acquisition in a distributed system, comprising the steps of:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system;

the manager system processing requests for the measurement variable values based on the predictive model on the manager system;

the disconnectable managed system determining that the predictive model is insufficiently accurate; and

the disconnectable managed system constructing and communicating an updated predictive model to the managed system, in response to said determining.

40. The method of claim 39, further comprising the step of the managed system disconnecting from the managed system except to update the predictive model.

41. A predictive model-based method for measurement data acquisition in a distributed system, comprising the steps of:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system;

the manager system processing requests for the measurement variable values based on the predictive model on the manager system;

the managed system disconnecting from the managed system;

the managed system measuring one or more actual values of the measurement variables and comparing the actual value against an accuracy bound;

if the actual values are within an accuracy bound associated with the predictive model on the managed system;

the managed system connecting and communicating a message to the manager system confirming that the actual values are within the accuracy bound; and

if the actual values are not within the accuracy bound associated with the predictive model on the managed system;

the managed system constructing an updated predictive model; and

connecting and communicating the updated predictive model to the manager system.

42. A predictive model-based method for measurement data acquisition in a distributed system, comprising the steps of:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system; and

the manager system processing requests for the measurement variable values based on the predictive model on the manager system;

wherein the system includes multiple managed systems that have one or more of different collection frequencies and unsynchronized clocks, said method further comprising the steps of;

said step of dynamically constructing including creating and associating a time stamp with a measurement variable value;

the manager system synchronizing data from the multiple managed systems, using the time stamp;

wherein said step of the manager system synchronizing data further comprises the step of adjusting a granularity of the time stamp using the predictive models.

43. A predictive model-based method for measurement data acquisition in a distributed system, comprising the steps of:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system; and

the manager system processing requests for the measurement variable values based on the predictive model on the manager system;

wherein the system includes multiple managed systems that have one or more of different collection frequencies and unsynchronized clocks, said method further comprising the steps of;

said step of dynamically constructing including creating and associating a time stamp with a measurement variable value;

the manager system synchronizing data from the multiple managed systems, using the time stamp;

wherein the unsynchronized clocks operate at a same rate, further comprising the step of;

the manager system predicting measurement variable values at sampling intervals based on a relative time stamp indexing a set of sampling intervals.

44. A predictive model-based method for measurement data acquisition in a distributed system, comprising the steps of:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system;

the manager system processing requests for the measurement variable values based on the predictive model on the manager system;

the disconnectable managed system monitoring measured values and using the predictive model on the disconnectable managed system to detect excessive deviations of predicted values from measured values;

the disconnectable managed system communicating to the manager system updates to the predictive model if said excessive deviations are detected; and

the manager system updating its predictive model based on the updates.

45. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for a predictive model-based method for measurement data acquisition in a distributed system, said method steps comprising:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system;

the manager system processing requests for the measurement variable values based on the predictive model on the manager system;

the disconnectable managed system determining that the predictive model is insufficiently accurate; and

the disconnectable managed system constructing and communicating an updated predictive model to the managed system, in response to said determining.

46. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for a predictive model-based method for measurement data acquisition in a distributed system, said method steps comprising:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system; and

the manager system processing requests for the measurement variable values based on the predictive model on the manager system;

wherein the system includes multiple managed systems that have one or more of different collection frequencies and unsynchronized clocks, said method further comprising the steps of;

said step of dynamically constructing including creating and associating a time stamp with a measurement variable value; and

the manager system synchronizing data from the multiple managed systems, using the time stamp;

wherein said step of the manager system synchronizing data further comprises the step of adjusting a granularity of the time stamp using the predictive models.

47. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for a predictive model-based method for measurement data acquisition in a distributed system, said method steps comprising:

dynamically constructing a predictive model, based on one or more non-static measurement variable values, on both a manager system and a disconnectable managed system; and

the manager system processing requests for the measurement variable values based on the predictive model on the manager system;

wherein the system includes multiple managed systems that have one or more of different collection frequencies and unsynchronized clocks, said method further comprising the steps of;

said step of dynamically constructing including creating and associating a time stamp with a measurement variable value; and

the manager system synchronizing data from the multiple managed systems, using the time stamp;

wherein the unsynchronized clocks operate at the same rate, further comprising the step of;

the manager system predicting measurement variable values at sampling intervals on a relative time stamp indexing a set of sampling intervals.

48. A predictive model-based measurement acquisition system for creating and managing one or more predictive models that predict values of measurement variables in distributed systems, said system comprising:

a manager system;

a disconnectable managed system;

means for dynamically constructing and coordinating a predictive model on both the disconnectable managed system and the manager system, the predictive model adapted for predicting values of said measurement variables; and

manager system means for processing requests for the measurement variable values based on the predictive model on the manager system;

the disconnectable managed system including;

agent memory means for storing the predictive model;

an agent data access facility means for providing measured values of said measurement variables; and

agent measurement repository means, coupled to the agent data access facility, for storing measurement variable values known to the manager; and

an agent model handler, coupled to the agent measurement repository, for updating and operating the predictive model on the disconnectable managed system;

the manager system including;

manager memory means for storing the predictive model;

manager measurement repository means for storing the measurement variable values in the memory;

one or more management applications, coupled to the memory, for requesting and using predicted values generated on the manager system and processing measurement variable values; and

manager model handler means, coupled to the memory, for operating the predictive model on the manager and updating the predictive model based on information communicated from the managed system.