Partitioned performance: adaptive predicted impact

• US 10,148,531 B1
• Filed: 09/30/2015
• Issued: 12/04/2018
• Est. Priority Date: 09/30/2015
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Inventors: Sledz, Dan, Walton, Jonathan, Powell, Daniel, Rang, Anton
• Primary Examiner(s): Meky, Moustafa M
• Assistant Examiner(s): Waugh, Thorne E
• Application Number:
  US14/870,159
• Time in Patent Office: 1,161 Days
• Field of Search: 709203-248
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: H04L 41/0893 Assignment of logical groups to network elements
• CPC Class Codes: H04L 41/0894 Policy-based network configuration management
• CPC Class Codes: H04L 43/0876 Network utilisation, e.g. volume of load or congestion level

Abstract

Implementations are provided herein for offering partitioned performance within a distributed file system and more specifically, for offering adaptive predicted impact of resource consumption by pending work items. Core resource consumption per work item can be estimated prior to processing the work item. When processing the work item, the actual amount of resources used to process the work item can be measured and recorded. The file system can then update future estimates for performing work items based on past results. Resources made available to process future requests can be throttled based on dynamically updated estimates of resource consumption by pending work items.

Claims

1. A method comprising:

identifying a first set of work items to be performed within a cluster of nodes operating as a distributed file system wherein a work item in the first set of work items is associated with a reference accounting token (“

RAT”

) specification among a set of RAT specifications;

generating a first file system resources estimate necessary to perform the first set of work items based at least in part on the associated RAT specifications among the set of RAT specifications;

processing the first set of work items by nodes within the cluster of nodes, wherein the processing the first set of work items includes generating a resource record for work items in the first set of work items, wherein the resource record contains an accounting of a set of resources used to process the work item, and wherein the set of resources includes at least one of a resource consumption time duration, network bandwidth, disk activity, central processing unit (“

CPU”

) usage, and memory usage; and

updating the first file system resources estimate based on the generating the resource record for work items in the first set of work items.

2. The method of claim 1, further comprising:

identifying a second set of work items to be performed within the cluster of nodes operating as the distributed file system, wherein a work item in the second set of work items is associated with a reference accounting token (“

RAT”

) specification among a set of RAT specifications;

generating a second file system resources estimate based on the updated first file system resources estimate and the associated RAT specifications among the set of RAT specifications;

processing the second set of work items by nodes within the cluster of nodes, wherein the processing the second set of work items includes generating a resource record for work items in the second set of work items; and

updating the second file system resources estimate based on the generating the resource record for work items in the second set of work items.

3. The method of claim 1, further comprising:

throttling a set of resources of the cluster of nodes available to process the first set of work items based on the first file system resources estimate, a throttling policy, and a set of outstanding work items to be performed.

4. The method of claim 2, further comprising:

throttling a set of resources of the cluster of nodes available to process the second set of work items based on the first file system resources estimate, the second file system resources estimate, a throttling policy, and a set of outstanding work items to be performed.

5. The method of claim 1, wherein the protocol is at least one of server message block, network file system, Hadoop distributed file system, file transfer protocol, object, hypertext transfer protocol, and a job engine of the distributed file system.

6. The method of claim 1, wherein the disk activity includes an indication of whether a disk cache was used to process the work item.

7. The method of claim 1, wherein network bandwidth includes a number of bytes sent, a number of bytes received, mbufs, mbuf clusters, and mbuf bytes.

8. A system comprising at least one storage device and at least one hardware processor configured to:

identify a first set of work items to be performed within a cluster of nodes operating as a distributed file system wherein a work item in the first set of work items is associated with a reference accounting token (“

RAT”

) specification among a set of RAT specifications;

generate a first file system resources estimate necessary to perform the first set of work items based at least in part on the associated RAT specifications among the set of RAT specifications;

process the first set of work items by nodes within the cluster of nodes, wherein the processing the first set of work items includes generating a resource record for work items in the first set of work items, wherein the resource record contains an accounting of a set of resources used to process the work item, and wherein the set of resources includes at least a resource consumption time duration, network bandwidth, disk activity, central processing unit (“

CPU”

) usage, and memory usage; and

update the first file system resources estimate based on the generating the resource record for work items in the first set of work items.

9. The system of claim 8 further configured to:

identify a second set of work items to be performed within the cluster of nodes operating as the distributed file system, wherein a work item in the second set of work items is associated with a reference accounting token (“

RAT”

) specification among a set of RAT specifications;

generate a second file system resources estimate based on the updated first file system resources estimate and the associated RAT specifications among the set of RAT specifications;

process the second set of work items by nodes within the cluster of nodes, wherein the processing the second set of work items includes generating a resource record for work items in the second set of work items; and

update the second file system resources estimate based on the generating the resource record for work items in the second set of work items.

10. The system of claim 8 further configured to:

throttle a set of resources of the cluster of nodes available to process the first set of work items based on the first file system resources estimate, a throttling policy, and a set of outstanding work items to be performed.

11. The system of claim 9 further configured to:

throttling a set of resources of the cluster of nodes available to process the second set of work items based on the first file system resources estimate, the second file system resources estimate, a throttling policy, and a set of outstanding work items to be performed.

12. The system of claim 8, wherein the protocol is at least one of server message block, network file system, Hadoop distributed file system, file transfer protocol, object, hypertext transfer protocol, and a job engine of the distributed file system.

13. The system of claim 8, wherein the disk activity includes an indication of whether a disk cache was used to process the work item.

14. A non-transitory computer readable medium with program instructions stored thereon to perform the following acts:

identifying a first set of work items to be performed within a cluster of nodes operating as a distributed file system wherein a work item in the first set of work items is associated with a reference accounting token (“

RAT”

) specification among a set of RAT specifications;

generating a first file system resources estimate necessary to perform the first set of work items based at least in part on the associated RAT specifications among the set of RAT specifications;

processing the first set of work items by nodes within the cluster of nodes, wherein the processing the first set of work items includes generating a resource record for work items in the first set of work items, wherein the resource record contains an accounting of a set of resources used to process the work item, and wherein the set of resources includes at least one of a resource consumption time duration, network bandwidth, disk activity, central processing unit (“

CPU”

) usage, and memory usage; and

updating the first file system resources estimate based on the generating the resource record for work items in the first set of work items.

15. The non-transitory computer readable medium of claim 14, with program instructions stored thereon to further perform the following acts:

identifying a second set of work items to be performed within the cluster of nodes operating as the distributed file system, wherein a work item in the second set of work items is associated with a reference accounting token (“

RAT”

) specification among a set of RAT specifications;

generating a second file system resources estimate based on the updated first file system resources estimate and the associated RAT specifications among the set of RAT specifications;

processing the second set of work items by nodes within the cluster of nodes, wherein the processing the second set of work items includes generating a resource record for work items in the second set of work items; and

updating the second file system resources estimate based on the generating the resource record for work items in the second set of work items.

16. The non-transitory computer readable medium of claim 14, with program instructions stored thereon to further perform the following acts:

throttling a set of resources of the cluster of nodes available to process the first set of work items based on the first file system resources estimate, a throttling policy, and a set of outstanding work items to be performed.

17. The non-transitory computer readable medium of claim 15, with program instructions stored thereon to further perform the following acts:

throttling a set of resources of the cluster of nodes available to process the second set of work items based on the first file system resources estimate, the second file system resources estimate, a throttling policy, and a set of outstanding work items to be performed.

18. The non-transitory computer readable medium of claim 14, wherein the protocol is at least one of server message block, network file system, Hadoop distributed file system, file transfer protocol, object, hypertext transfer protocol, and a job engine of the distributed file system.

19. The non-transitory computer readable medium of claim 14, wherein the disk activity includes an indication of whether a disk cache was used to process the work item.

20. The non-transitory computer readable medium of claim 14, wherein network bandwidth includes a number of bytes sent, a number of bytes received, mbufs, mbuf clusters, and mbuf bytes.