Virtual splitter

• US 9,507,845 B1
• Filed: 03/27/2014
• Issued: 11/29/2016
• Est. Priority Date: 03/27/2014
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Inventors: Natanzon, Assaf, Cohen, Saar
• Primary Examiner(s): NGUYEN, CAM LINH T
• Application Number:
  US14/227,058
• Time in Patent Office: 978 Days
• Field of Search: 707/634
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: G06F 11/1474 in transactions G06F16/20 takes precedence
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 2201/815 Virtual
• CPC Class Codes: G06F 2201/82 Solving problems relating to consistency
• CPC Class Codes: G06F 9/45533 Hypervisors; Virtual machine monitors
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

A system, program product, and computer implemented method for replicating a consistency group comprising monitoring the latency between one or more splitters of the consistency group and replication appliances in the replication cluster for the consistency group (CG); wherein each replication appliance of the replication appliances is configured to be able to receive IO from the one or more splitters, determining which replication appliance of the replication appliances has the lowest latency after including additional latency resulting from assignment of the CG to the replication appliance, and configuring the splitter to replicate IO from the CG to the replication appliance determined to have the lowest latency.

Claims

1. A computer implemented method for replicating a consistency group comprising:

monitoring input/output (IO) latency between one or more splitters of the consistency group and replication appliances in a replication cluster for the consistency group (CG);

wherein each replication appliance of the replication appliances is configured to be able to receive IO from the one or more splitters;

wherein the one or more splitters split IO at a production site by intercepting IO in a data path being written to storage, creating a copy of the IO, and sending the copy to one or more of the replication appliances;

wherein the IO latency is the time to send the copy of the IO to a replication appliance;

determining which replication appliance of the replication appliances has a lowest latency to receive IO split from a splitter of the one or more splitters;

wherein the lowest latency includes the IO latency and an additional latency resulting from a proposed assignment of the CG split by the splitter to the replication appliance;

wherein the additional latency results from an addition load on the replication appliance handling IO split from the splitter of the one or more splitter; and

configuring the splitter to replicate IO from the CG to the replication appliance determined to have the lowest latency.

2. The method of claim 1 wherein each replication appliance in the replication cluster runs on a different hypervisor;

wherein a hypervisor is enabled to run one or more virtual machines.

3. The method of claim 2 wherein at least one of the hypervisors is located on a geographically disparate site from at least one other hypervisor.

4. The method of claim 2 wherein the monitoring is performed by a master replication appliance and wherein the master replication appliance determines and assigns splitters to replication appliances in the replication cluster.

5. The method of claim 3 wherein the splitter intercepts IO from a virtual machine running in a hypervisor being sent to a virtual disk of the virtual machine.

6. The method of claim 1 wherein the current load on each replication appliance is considered before the determining and wherein at least one splitter of the one or more slitters runs in a virtualization layer in a hypervisor.

7. The method of claim 5 wherein the best allocation of CGs to replication appliances is determined by a machine learning algorithm.

21. The method of claim 6 wherein the monitoring, determining, and configuring are performed periodically.

8. A computer program product comprising:

a non-transitory computer readable medium encoded with computer executable program code for replication of data, the code configured to enable the execution of;

monitoring input/output (IO) latency between one or more splitters of a consistency group and replication appliances in a replication cluster for the consistency group (CG);

wherein each replication appliance of the replication appliances is configured to be able to receive IO from the one or more splitters;

wherein the one or more splitters split IO at a production site by intercepting IO in a data path being written to storage, creating a copy of the IO, and sending the copy to one or more of the replication appliances;

wherein the IO latency is the time to send the copy of the IO to a replication appliance;

determining which replication appliance of the replication appliances has a lowest latency to receive IO split from a splitter of the one or more splitters;

wherein the lowest latency includes the IO latency and an additional latency resulting from a proposed assignment of the CG split by the splitter to the replication appliance;

wherein the additional latency results from an addition load on the replication appliance handling IO split from the splitter of the one or more splitter; and

configuring the splitter to replicate IO from the CG to the replication appliance determined to have the lowest latency.

9. The computer program product of claim 8 wherein each replication appliance in the replication cluster runs on a different hypervisor;

wherein a hypervisor is enabled to run one or more virtual machines.

10. The computer program product of claim 8 wherein at least one of the hypervisors is located on a geographically disparate site from at least one other hypervisor.

11. The computer program product of claim 8 wherein the monitoring is performed by a master replication appliance and wherein the master replication appliance determines and assigns splitters to replication appliances in the replication cluster.

12. The computer program product of claim 9 wherein the splitter intercepts IO from a virtual machine running in a hypervisor being sent to a virtual disk of the virtual machine.

13. The computer program product of claim 8 wherein the best allocation of CGs to replication appliances is determined by a machine learning algorithm.

14. The computer program product of claim 8 wherein the current load on each appliance is considered before the determining and wherein at least one splitter of the one or more slitters runs in a virtualization layer in a hypervisor.

15. A system for data replication, the system comprising:

a production site;

a replication cluster; and

computer-executable logic operating in memory, wherein the computer-executable program logic is configured for execution of;

monitoring input/output (IO) latency between one or more splitters of a consistency group and replication appliances in the replication cluster for the consistency group (CG);

wherein each replication appliance of the replication appliances is configured to be able to receive IO from the one or more splitters;

wherein the one or more splitters split IO at the production site by intercepting IO in a data path being written to storage, creating a copy of the IO, and sending the copy to one or more of the replication appliances;

wherein the IO latency is the time to send the copy of the IO to a replication appliance;

determining which replication appliance of the replication appliances has a lowest latency to receive IO split from a splitter of the one or more splitters;

wherein the lowest latency includes the IO latency and an additional latency resulting from a proposed assignment of the CG split by the splitter to the replication appliance;

wherein the additional latency results from an addition load on the replication appliance handling IO split from the splitter of the one or more splitter; and

configuring the splitter to replicate IO from the CG to the replication appliance determined to have the lowest latency.

16. The system of claim 15 wherein each replication appliance in the replication cluster runs on a different hypervisor;

wherein a hypervisor is enabled to run one or more virtual machines.

17. The system of claim 15 wherein at least one of the hypervisors is located on a geographically disparate site from at least one other hypervisor.

18. The system of claim 15 wherein the monitoring is performed by a master replication appliance and wherein the master replication appliance determines and assigns splitters to replication appliances in the replication cluster.

19. The system of claim 15 wherein the best allocation of CGs to replication appliances is determined by a machine learning algorithm.

20. The system of claim 15 wherein the current load on each replication appliance is considered before the determining and wherein at least one splitter of the one or more slitters runs in a virtualization layer in a hypervisor.