TUNING AND OPTIMIZING DISTRIBUTED SYSTEMS WITH DECLARATIVE MODELS

• US 20110179151A1
• Filed: 03/31/2011
• Published: 07/21/2011
• Est. Priority Date: 06/29/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Zunino, Gilles C.J.A., Panditharadhya, Prasad S., Borsa, Mariusz G., Sedukhin, Igor, Eshner, Daniel, Kulkarni, Amol S.
• Granted Patent: US 8,099,494 B2
• Time in Patent Office: Days
• US Class Current: 709/221
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 2201/86 Event-based monitoring

Abstract

A system for automatically adjusting operation of a distributed application program includes analytics means having a monitoring component. The monitoring component receives one or more event streams of data corresponding to execution of one or more modules of the distributed application program. The monitoring component joins the received event streams with one or more declarative models to create operational data. A forensics component of the analytics means queries, such as by using data from a declarative model store, the operational data to identify trends or behavior information for the various modules or components of the distributed application program. A planner component then reviews data created by the forensics component to identify any needed changes to the declarative models. The planner component passes any modified declarative models back to the system, so that the distributed application program can operate more efficiently using the changes to the declarative models, as needed.

Claims

1. A method of automatically monitoring and adjusting behaviors of a distributed application program through management of one or more declarative models, comprising:

sending an initial set of instructions corresponding to at least one declarative model of a distributed application program to at least one platform-specific driver of an implementation means of a server system and which causes the at least one platform-specific driver to responsively provide instructions for executing the one or more modules as part of the distributed application program;

after initiation of execution of the distributed application program, receiving one or more event streams through the at least one platform-specific driver, the one or more received event streams containing information about real-time performance of the distributed application program; and

modifying the at least one declarative model in accordance with the received execution information, wherein the modification to the at least one declarative model adjusts one or more intended operations of one or more modules of the distributed application program.

2. The method as recited in claim 1, wherein the one or more event streams are received in response to a query that is scheduled to occur periodically.

3. The method as recited in claim 1, further comprising an act of a monitoring component creating operational data from the received one or more event streams, wherein the operational data includes behavior information metrics about execution of the distributed application program.

4. The method as recited in claim 3, wherein the act of creating operational data further comprises joining data of at least one of the one or more event streams with data in an declarative model repository.

5. The method as recited in claim 3, further comprising an act of a forensics component querying the operational data with one or more queries to identify one or more operational trends of the distributed application program.

6. The method as recited in claim 5, wherein at least one of the one or more queries includes one or more identifiers of a module in the distributed application program.

7. The method as recited in claim 5, wherein the one or more operational trends include an indication that one or more hardware components on which the distributed application program is executing are operating incorrectly or are failing.

8. The method as recited in claim 5, wherein the one or more operational trends include an indication that one or more software modules of the distributed application program are executing incorrectly or are failing.

9. The method as recited in claim 5, wherein the one or more operational trends include an indication of a frequency with which a machine that is executing at least a portion of the distributed application program is rebooting or otherwise restarting.

10. The method as recited in claim 5, further comprising an act of a planner component receiving the one or more operational trends from the forensics component.

11. The method as recited in claim 10, further comprising an act of the planner component correlating the one or more operational trends received from the forensics component with data of the at least one declarative model.

12. The method as recited in claim 11, wherein the act of modifying the at least one declarative model in accordance with the received execution information further comprises an act of the planner component modifying the at least one declarative model to create a modified declarative model.

13. The method as recited in claim 12, wherein the modified declarative model identifies a different set of one or more hardware components on which at least a portion of the distributed application program should be executed.

14. The method as recited in claim 12, wherein the modified declarative model reflects a change in execution or use policy for at least a portion of the distributed application program.

15. The method as recited in claim 12, further comprising an act of sending the modified declarative model to the implementation means, wherein information of the modified declarative model can be processed through the platform-specific driver.

16. The method as recited in claim 15, further comprising an act of the implementation means sending a subsequent set of instructions corresponding to the modified declarative model to the platform-specific driver.

17. A computing system comprising:

at least one processor;

one or more computer storage medium having stored instructions which, when executed by the at least one processor, implement a method of automatically monitoring and adjusting behaviors of a distributed application program through management of one or more declarative models, comprising;

providing to at least one platform-specific driver of a server system an initial set of instructions corresponding to at least one declarative model of a distributed application program and which causes the at least one platform-specific driver to responsively provide instructions for executing the one or more modules as part of the distributed application program;

receiving one or more event streams through the at least one platform-specific driver after initiated execution of the distributed application program, the one or more received event streams containing information about real-time performance of the distributed application program; and

modifying the at least one declarative model in accordance with the received execution information, wherein the modification to the at least one declarative model adjusts one or more intended operations of one or more modules of the distributed application program.

18. A computer storage product having computer-executable instructions stored thereon that, when executed, cause one or more processors to perform a method of automatically monitoring and adjusting behaviors of a distributed application program through management of one or more declarative models, comprising the acts of:

sending an initial set of instructions corresponding to at least one declarative model of a distributed application program to at least one platform-specific driver of implementation means, the at least one declarative model describing intent for operation of one or more modules of the distributed application program, and which result in the at least one platform-specific driver providing instructions for executing the one or more modules as part of the distributed application program;

after execution of the distributed application program has been initiated, receiving one or more event streams through the at least one platform-specific driver, the one or more event streams received through the at least one platform-specific driver containing information about real-time performance of the distributed application program during execution of the distributed application program within the computerized environment; and

modifying the at least one declarative model in accordance with the received execution information, wherein the modification to the at least one declarative model adjusts one or more intended operations of one or more modules of the distributed application program.