Granted Patent B1

US 11,330,078 · App. 17/386,402 · Granted May 10, 2022

Inventors: Shelesh Chopra (Bangalore, IN); Hemant Ramesh Gaikwad (Bangalore, IN); Rahul Deo Vishwakarma (Bangalore, IN)

Assignee: EMC IP Holding Company LLC

H04L67/32

Abstract

Claims (91)

1. A method for managing updates of a data manager, the method comprising:

identifying, by a distributed control manager of a data manager, a service update event, wherein the data manager comprises services;

in response to identifying the service update event:

identifying a service sub-tree based on a service call graph;

generating an update sequence for the service sub-tree using the service call graph;

predicting an update window for the service sub-tree;

selecting a first service of the service sub-tree based on the update sequence, wherein the first service comprises a first standby service instance and a first active service instance;

generating a backup of a first portion of a services shared data volume repository associated with the first service;

applying a first portion of an update to the first standby service instance to obtain a first updated active service instance;

making a first determination that a first performance and reliability check associated with the first updated active service instance is below a threshold; and

in response to the first determination:

applying the first portion of the update to a second standby service instance, wherein the second standby service instance comprises the first active service instance.

2. The method of claim 1 , wherein applying the first portion of the update to the first standby service instance comprises changing the first active service instance to the second standby service instance.

3. The method of claim 2 , wherein the service call graph specifies a hierarchical relationship between the services.

4. The method of claim 3 , wherein the update sequence specifies an order in which the services of the service sub-tree is to be updated.

5. The method of claim 1 , wherein predicting the update window for the service sub-tree comprises:

predicting service sub-tree update completion time;

predicting future requests for services associated with the service sub-tree to obtain a future request density transition set;

performing peak smoothing on the future request density transition set;

performing continuous accuracy improvement to obtain an updated future request density transition set;

identifying dynamic update window identification markings (DUWIMs) for the service sub-tree based on the updated future request density transition set; and

identifying the update window using the DUWIMs.

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

after applying the first portion of the update to the second standby service instance:

selecting a second service of the service sub-tree based on the update sequence, wherein the second service comprises a third standby service instance and a second active service instance;

generating a backup of a second portion of a services shared data volume repository associated with the second service;

applying a second portion of an update to the third standby service instance to obtain a second updated active service instance;

making a second determination that a second performance and reliability check associated with the second updated active service instance is above a threshold; and

in response to the second determination:

reversing the second portion of the update to the second updated active service instance to obtain the third standby service instance.

7. The method of claim 1 , the method further comprising:

prior to the identifying a service update event:

identifying, by a discovery manager of the data manager, a service change event; and

updating a service routing table based on the service change event.

8. The method of claim 7 , wherein the service update event comprises one selected from a group consisting of:

obtaining a service registration request; and

a timeout of a service.

9. A non-transitory computer readable medium comprising computer readable program code, which when executed by a computer processor enables the computer processor to perform a method for managing updates of a data manager, the method comprising:

identifying, by a distributed control manager of a data manager, a service update event, wherein the data manager comprises services;

in response to identifying the service update event:

identifying a service sub-tree based on a service call graph;

generating an update sequence for the service sub-tree using the service call graph;

predicting an update window for the service sub-tree;

selecting a first service of the service sub-tree based on the update sequence, wherein the first service comprises a first standby service instance and a first active service instance;

generating a backup of a first portion of a services shared data volume repository associated with the first service;

applying a first portion of an update to the first standby service instance to obtain a first updated active service instance;

making a first determination that a first performance and reliability check associated with the first updated active service instance is below a threshold; and

in response to the first determination:

applying the first portion of the update to a second standby service instance, wherein the second standby service instance comprises the first active service instance.

10. The non-transitory computer readable medium of claim 9 , wherein applying the first portion of the update to the first standby service instance comprises changing the first active service instance to the second standby service instance.

11. The non-transitory computer readable medium of claim 10 , wherein the service call graph specifies a hierarchical relationship between the services.

12. The non-transitory computer readable medium of claim 11 , wherein the update sequence specifies an order in which the services of the service sub-tree is to be updated.

13. The non-transitory computer readable medium of claim 9 , wherein predicting the update window for the service sub-tree comprises:

predicting service sub-tree update completion time;

predicting future requests for services associated with the service sub-tree to obtain a future request density transition set;

performing peak smoothening on the future request density transition set;

performing continuous accuracy improvement to obtain an updated future request density transition set;

identifying dynamic update window identification markings (DUWIMs) for the service sub-tree based on the updated future request density transition set; and

identifying the update window using the DUWIMs.

14. The non-transitory computer readable medium of claim 9 , the method further comprising:

after applying the first portion of the update to the second standby service instance:

selecting a second service of the service sub-tree based on the update sequence, wherein the second service comprises a third standby service instance and a second active service instance;

generating a backup of a second portion of a services shared data volume repository associated with the second service;

applying a second portion of an update to the third standby service instance to obtain a second updated active service instance;

making a second determination that a second performance and reliability check associated with the second updated active service instance is above a threshold; and

in response to the second determination:

reversing the second portion of the update to the second updated active service instance to obtain the third standby service instance.

15. The non-transitory computer readable medium of claim 9 , the method further comprising:

prior to the identifying a service update event:

identifying, by a discovery manager of the data manager, a service change event; and

updating a service routing table based on the service change event.

16. The non-transitory computer readable medium of claim 15 , wherein the service update event comprises one selected from a group consisting of:

obtaining a service registration request; and

a timeout of a service.

17. A system for managing updates of a data manager, the system comprising:

a data manager comprising services;

a distributed control manager of the data manager, comprising a processor and memory, configured to:

identify a service update event;

in response to identifying the service update event:

identify a service sub-tree based on a service call graph;

generate an update sequence for the service sub-tree using the service call graph;

predict an update window for the service sub-tree;

select a first service of the service sub-tree based on the update sequence, wherein the first service comprises a first standby service instance and a first active service instance;

generate a backup of a first portion of a services shared data volume repository associated with the first service;

apply a first portion of an update to the first standby service instance to obtain a first updated active service instance;

make a first determination that a first performance and reliability check associated with the first updated active service instance is below a threshold; and

in response to the first determination:

apply the first portion of the update to a second standby service instance, wherein the second standby service instance comprises the first active service instance.

18. The system of claim 17 , wherein applying the first portion of the update to the first standby service instance comprises changing the first active service instance to the second standby service instance.

19. The system of claim 18 , wherein the service call graph specifies a hierarchical relationship between the services.

20. The system of claim 19 , wherein the update sequence specifies an order in which the services of the service sub-tree is to be updated.

Assignments (8)

Priority Claims (1)