MANAGING NODES IN A HIGH-PERFORMANCE COMPUTING SYSTEM USING A NODE REGISTRAR

• US 20120324456A1
• Filed: 06/16/2011
• Published: 12/20/2012
• Est. Priority Date: 06/16/2011
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Teather, Gregory Wray
• Application Number:
  US13/162,130
• Publication Number:
  US 20120324456A1
• Time in Patent Office: Days
• US Class Current: 718/102
• CPC Class Codes: G06F 2209/505 Clust
• CPC Class Codes: G06F 9/465 Distributed object oriented systems remote method invocation [RMI] G06F9/548
• CPC Class Codes: G06F 9/4843 by program, e.g. task dispatcher, supervisor, operating system
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers

Abstract

A method of managing nodes in a high-performance computing (HPC) system, which includes a management subsystem and a job scheduler subsystem, includes providing a node registrar subsystem. Logical node management functions are performed with the node registrar subsystem. Other management functions are performed with the management subsystem using the node registrar subsystem. Job scheduling functions are performed with the job scheduler subsystem using the node registrar subsystem.

Claims

1. A method of managing nodes in a high-performance computing (HPC) system, which includes a management subsystem and a job scheduler subsystem, the method comprising:

providing a node registrar subsystem;

performing logical node management functions with the node registrar subsystem;

performing other management functions with the management subsystem using the node registrar subsystem; and

performing job scheduling functions with the job scheduler subsystem using the node registrar subsystem.

2. The method of claim 1, wherein the management subsystem and the job scheduler subsystem are each a client of the node registrar subsystem.

3. The method of claim 1, wherein the node registrar subsystem comprises a stateless node registrar service and a database for storing node information for the nodes in the HPC system.

4. The method of claim 3, and further comprising:

accessing the stored node information with the management subsystem using the node registrar service.

5. The method of claim 3, and further comprising:

updating node properties in the database with the management subsystem using the node registrar service.

6. The method of claim 3, and further comprising:

query the nodes by property with the management subsystem using the node registrar service.

7. The method of claim 3, and further comprising:

query the nodes by group with the management subsystem using the node registrar service.

8. The method of claim 3, and further comprising:

accessing the stored node information with the job scheduler subsystem using the node registrar service.

9. The method of claim 3, and further comprising:

updating node properties in the database with the job scheduler subsystem using the node registrar service.

10. The method of claim 3, and further comprising:

query the nodes by property with the job scheduler subsystem using the node registrar service.

11. The method of claim 3, and further comprising:

query the nodes by group with the job scheduler subsystem using the node registrar service.

12. The method of claim 3, wherein the database includes a node table with each row in the node table corresponding to one of the nodes in the HPC system, and each column listing properties of the nodes in the HPC system.

13. The method of claim 1, wherein the logical node management functions performed by the node registrar subsystem include adding nodes and removing nodes.

14. The method of claim 13, wherein the logical node management functions performed by the node registrar subsystem further include updating node properties and handling state transitions of nodes.

15. The method of claim 14, wherein the logical node management functions performed by the node registrar subsystem further include grouping nodes.

16. The method of claim 13, and further comprising:

performing node deployment, node configuration management, infrastructure configuration management, and node monitoring with the management subsystem.

17. A computer-readable storage medium storing computer-executable instructions that when executed by at least one processor cause the at least one processor to perform a method of managing nodes in a high-performance computing (HPC) system, wherein the HPC system includes a management subsystem and a job scheduler subsystem, the method comprising:

performing logical node management functions with a node registrar subsystem;

performing other management functions with the management subsystem using the node registrar subsystem; and

performing job scheduling functions with the job scheduler subsystem using the node registrar subsystem.

18. The computer-readable storage medium of claim 17, wherein the node registrar subsystem comprises a stateless node registrar service and a database for storing node information for the nodes in the HPC system, and wherein the management subsystem and job scheduler subsystem are configured to access the database using the node registrar service.

19. The computer-readable storage medium of claim 17, wherein the logical node management functions performed by the node registrar subsystem include adding nodes, removing nodes, updating node properties, and handling state transitions of nodes, and wherein the other management functions performed by the management subsystem include node deployment, node configuration management, infrastructure configuration management, and node monitoring.

20. A method of managing nodes in a high-performance computing (HPC) system, which includes a management subsystem and a job scheduler subsystem, the method comprising:

providing a node registrar subsystem, wherein the node registrar subsystem comprises a stateless node registrar service and a database for storing node information for the nodes in the HPC system;

performing logical node management functions with the node registrar subsystem;

performing other management functions with the management subsystem using the node registrar service to access the database; and

performing job scheduling functions with the job scheduler subsystem using the node registrar service to access the database.