System and method for load balancing with XOR star and XOR chain
The disclosure relates to data protection management for geographically distributed storage systems. Specifically, the present disclosure provides a mechanism for load balancing by combining the techniques of XOR Star and XOR Chain to control the protection of data internally. The load balancing is done by each zone without the need for an external load balancer. In addition, the load balancing mechanism provides the ability for native (e.g. without additional configuration) support of disaster recovery.
1. A method of performing load balancing within a data storage system, comprising:
storing chunks of data amongst a set of zones within the data storage system;
performing, by at least a first and second zone, a first mode of replication, wherein the first mode of replication includes
selecting a first remote zone based on a policy, and
replicating one of its own chunks to the first remote zone, wherein the first remote zone creates a first XOR chunk to store locally from the replicated chunks received from the first and second zones;
determining, by the first or second zone, a load imbalance of a target zone; and
in response to determining the load imbalance, performing, by the first or second zone, a second mode of replication, wherein the second mode of replication includes
creating a second XOR chunk from one of its own chunks and a replicated chunk received from a second remote zone, and
replicating the second XOR chunk to the target zone to rebalance a load amongst the set of zones.
2. The method of claim 1 , further comprising determining, by the first or second zone, a load rebalance of the target zone, and in response, switching back to the first mode of replication.
3. The method of claim 1 , wherein creating the second XOR chunk includes creating a partial XOR chunk when the load imbalance includes at least two target zones.
4. The method of claim 1 , wherein performing the first mode of replication is repeated by the first and second zones until the load imbalance is determined.
5. The method of claim 1 , wherein the policy for selecting the first remote zone is a round-robin policy.
6. The method of claim 1 , wherein selecting the first remote zone is performed individually by the first and second zones and without coordination from an external load balancer.
7. A data storage system, comprising:
a memory storing instructions; and
one or more processors coupled to the memory to execute the instructions from the memory, the one or more processors being configured to perform operations, the operations comprising:
storing chunks of data amongst a set of zones within the data storage system;
performing, by at least a first and second zone, a first mode of replication, wherein the first mode of replication includes
selecting a first remote zone based on a policy, and
replicating one of its own chunks to the first remote zone, wherein the first remote zone creates a first XOR chunk to store locally from the replicated chunks received from the first and second zones;
determining, by the first or second zone, a load imbalance of a target zone; and
in response to determining the load imbalance, performing, by the first or second zone, a second mode of replication, wherein the second mode of replication includes
creating a second XOR chunk from one of its own chunks and a replicated chunk received from a second remote zone, and
replicating the second XOR chunk to the target zone to rebalance a load amongst the set of zones.
8. The system of claim 7 , the operations further comprising determining, by the first or second zone, a load rebalance of the target zone, and in response, switching back to the first mode of replication.
9. The system of claim 7 , wherein creating the second XOR chunk includes creating a partial XOR chunk when the load imbalance includes at least two target zones.
10. The system of claim 7 , wherein performing the first mode of replication is repeated by the first and second zones until the load imbalance is determined.
11. The system of claim 7 , wherein the policy for selecting the first remote zone is a round-robin policy.
12. The system of claim 7 , wherein selecting the first remote zone is performed individually by the first and second zones and without coordination from an external load balancer.
13. A non-transitory machine-readable medium storing instructions which, when executed by one or more processors of a computing device, cause the computing device to perform operations comprising:
storing chunks of data amongst a set of zones within the data storage system;
performing, by at least a first and second zone, a first mode of replication, wherein the first mode of replication includes
selecting a first remote zone based on a policy, and
replicating one of its own chunks to the first remote zone, wherein the first remote zone creates a first XOR chunk to store locally from the replicated chunks received from the first and second zones;
determining, by the first or second zone, a load imbalance of a target zone; and
in response to determining the load imbalance, performing, by the first or second zone, a second mode of replication, wherein the second mode of replication includes
creating a second XOR chunk from one of its own chunks and a replicated chunk received from a second remote zone, and
replicating the second XOR chunk to the target zone to rebalance a load amongst the set of zones.
14. The medium of claim 13 , further comprising determining, by the first or second zone, a load rebalance of the target zone, and in response, switching back to the first mode of replication.
15. The medium of claim 13 , wherein creating the second XOR chunk includes creating a partial XOR chunk when the load imbalance includes at least two target zones.
16. The medium of claim 13 , wherein performing the first mode of replication is repeated by the first and second zones until the load imbalance is determined.
17. The medium of claim 13 , wherein the policy for selecting the first remote zone is a round-robin policy.
18. The medium of claim 13 , wherein selecting the first remote zone is performed individually by the first and second zones and without coordination from an external load balancer.