Progressively implementing declarative models in distributed systems

• US 8,239,505 B2
• Filed: 06/29/2007
• Issued: 08/07/2012
• Est. Priority Date: 06/29/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Sedukhin, Igor, Eshner, Daniel, Kulkarni, Amol S., Venkataramanappa, Girish M., Vannelli, Leo S. III, Mohanty, Sumit, Sahi, Sundeep
• Primary Examiner(s): STRANGE, AARON N
• Application Number:
  US11/771,816
• Publication Number:
  US 20090006062A1
• Time in Patent Office: 1,866 Days
• Field of Search: 709/220, 709/221, 709/223, 709/224, 707/100, 717/174
• US Class Current: 709/221
• CPC Class Codes: G06F 8/10 Requirements analysis; Specification techniques
• CPC Class Codes: G06F 8/35 model driven
• CPC Class Codes: G06F 8/64 Retargetable

Abstract

A system for automatically implementing high-level instructions in a distributed application program, where the high-level instructions reflect the behavior of the distributed application program, includes at least a tools component. The tools component is used to write high-level instructions in the form of declarative models, and place them in a repository. An executive component then receives the declarative models from the repository and refines them (e.g., via progressive elaboration) until there are no ambiguities. A platform-specific driver then translates the commands from the executive component, effectively turning the declarative model instructions into a set of imperative actions to be implemented in one or more application containers. The platform-specific driver also relays one or more event streams to an analytics means, which can result in modifications to the declarative models and corresponding new sets of instructions coming through the platform-specific driver at a later point.

Claims

1. At a server having one or more implementations means and one or more analytics means configured in a distributed computerized system environment to manage one or more distributed application programs through one or more platform-specific drivers, a method performed by a processor of the server of automatically implementing one or more sets of high-level instructions in a distributed application program during execution thereof using declarative models, comprising the acts of:

storing, by the processor, one or more declarative models in a repository, the one or more declarative models including high-level instructions regarding one or more operations of a distributed application program;

refining the one or more declarative models to include contextual information regarding operations of the distributed application program;

translating the one or more refined declarative models into sets of one or more platform-specific commands to be implemented by one or more application containers;

sending the translated commands to the one or more application containers, wherein the translated commands are received and executed by the one or more application containers to configure and control the distributed application program;

while the distributed application program is executing, receiving one or more modifications to the one or more declarative models in the repository; and

in response to the one or more modifications, refining and translating the one or more modified declarative models into sets of one or more platform-specific commands, and sending the sets of one or more platform-specific commands to the one or more application containers to be executed by the one or more application containers to implement the distributed application program such that the one or more modifications to the one or more declarative models are reflected in the distributed application program.

2. The method as recited in claim 1, wherein the one or more declarative models describe one or more distributed application program operational intents, and do not include specific information for implementing the distributed application program operational intents.

3. The method as recited in claim 1, further comprising, upon identifying the one or more modifications, an executive component scheduling a refinement process that adds information to a corresponding modified form of the one or more declarative models.

4. The method as recited in claim 1, wherein refining the one or more declarative models further comprises an executive component adding information regarding dependencies between elements in any of the one or more declarative models and inventory of the one or more application containers.

5. The method as recited in claim 1, wherein refining the one or more declarative models further comprises adding any missing data in the one or more declarative models that is required for translation by a platform-specific driver.

6. The method as recited in claim 5, wherein adding any missing data includes identifying specific computer systems on which one or more modules of the distributed application program should run.

7. The method as recited in claim 5, wherein adding any missing data includes determining one or more transport protocols to use between connected modules used by the distributed application program.

8. The method as recited in claim 5, wherein adding any missing data includes identifying any dependent data within at least one of the one or more declarative models.

9. The method as recited in claim 8, further comprising the acts of:

identifying one or more URI addresses for the identified dependent data; and

propagating any additional URI addresses between the identified dependent data.

10. The method as recited in claim 5, wherein adding any missing data further comprises identifying any constraints in the one or more declarative models regarding location of one or more distributed application program modules on a particular computer system.

11. The method as recited in claim 5, wherein adding any missing data further comprises completing the one or more declarative models so that they can be translated by a platform-specific driver.

12. The method as recited in claim 11, wherein completing the one or more declarative models to be translated further comprises including an address of a platform -specific driver in each of the refined, completed one or more declarative models.

13. The method as recited in claim 12, further comprising the executive component verifying the one or more declarative models with each platform-specific driver to determine that each of the one or more declarative models can be translated.

14. At a server having one or more implementations means and one or more analytics means configured in a distributed computerized system environment to manage one or more distributed application programs through one or more platform-specific drivers, a method performed by a processor of the server of automatically implementing one or more sets of high-level instructions in a distributed application program during execution thereof in one or more distributed application containers using declarative models, comprising the acts of:

receiving a set of new one or more declarative models from a repository, the new one or more declarative models including high-level instructions regarding operations of a distributed application program;

implementing the distributed application program through an implementation means and one or more application containers by generating a first set of low-level platform-specific commands from the new one or more declarative models and sending the first set of low-level platform-specific commands to the one or more application containers to be executed;

modifying the new one or more declarative models in the repository to address performance information of the distributed application program that is received from the one or more application containers while the distributed application program is executing;

identifying the modification to the new one or more declarative models in the repository via one or more analytics means; and

implementing an updated version of the distributed application program through the implementation means and the one or more application containers by generating a second set of low-level platform-specific commands from the modified new one or more declarative models and sending the second set of low-level platform-specific commands to the one or more application containers to be executed.

15. The method as recited in claim 14, wherein implementing the new one or more declarative models further comprises:

preparing a refined set of one or more new instructions corresponding to the new one or more declarative models; and

verifying that the set of one or more new instructions can be translated by a platform -specific driver into the first set of low-level commands.

16. The method as recited in claim 14, wherein implementing the updated version of the distributed application program further comprises:

preparing a new refined set of one or more updated instructions corresponding to the modified new one or more declarative models; and

verifying that the new refined set of one or more updated instructions can be translated by a platform-specific driver into the second set of low-level commands.

17. The method as recited in claim 14, wherein the performance information is received from the one or more application containers in one or more event streams.

18. The method as recited in claim 17, wherein the one or more event streams are received via a platform-specific driver that interfaces with the one or more application containers.

19. At a server having one or more implementations means and one or more analytics means configured in a distributed computerized system environment to manage one or more distributed application programs through one or more platform-specific drivers, a computer program storage product comprising computer executable instructions stored thereon that, when executed, cause one or more processors at the server to perform a method comprising the acts of:

storing, by the processor, one or more declarative models in a repository, the one or more declarative models including high-level instructions regarding one or more operations of a distributed application program;

refining the one or more declarative models to include contextual information regarding operations of the distributed application program;

translating the one or more refined declarative models into sets of one or more platform -specific commands to be implemented by one or more application containers;

sending the translated commands to the one or more application containers, wherein the translated commands are received and executed by the one or more application containers to configure and control the distributed application program;

while the distributed application program is executing, receiving one or more modifications to the one or more declarative models in the repository; and

in response to the one or more modifications, refining and translating the one or more modified declarative models into sets of one or more platform-specific commands, and sending the sets of one or more platform-specific commands to the one or more application containers to be executed by the one or more application containers to implement the distributed application program such that the one or more modifications to the one or more declarative models are reflected in the distributed application program.