Multi-cluster warehouse

1. A method for implementing a fault-tolerant data warehouse using availability zones, comprising:

allocating a plurality of compute groups each comprising a plurality of compute resources as part of a data warehouse, wherein each of the compute groups accesses data using a plurality of queries with one or more databases in one or more cloud storage resources, wherein each of the compute groups includes a plurality of processing resources;

routing, by a processing device, one or more of the plurality of queries to each of the plurality of compute groups, the plurality of compute groups comprising at least two processing units in different availability zones, an availability zone comprising one or more data centers, each data center comprising redundant power, networking, and connectivity, the one or more of the plurality of queries having a common session identifier with a query previously provided to the compute group, the compute group further determined to be caching a data segment usable by the one or more of the plurality of queries, wherein in response to the one or more of the plurality of queries, each of the plurality of compute groups performs a database operation on a particular portion of a database table; and

changing a total number of the plurality of compute groups using at least a comparison of a targeted degree of parallelism and a runtime computed degree of parallelism of the runtime computed degree of parallelism is computed using a number of queries running at an input degree of parallelism and a change is at least one of creating a new compute group or deleting an existing compute group.

2. The method of claim 1 , wherein deleting the existing compute group comprises removing the compute group down to a predetermined minimum number of compute groups.

3. The method of claim 1 , wherein deleting the existing compute group comprises:

determining whether a current workload is serviceable by fewer compute groups than the plurality of compute groups while meeting a performance metric; and

decommissioning at least one compute group of the plurality of compute groups in response to determining that the current workload is serviceable by fewer than the plurality of compute groups.

4. The method of claim 1 , wherein deleting the existing compute group comprises:

determining whether a current workload is serviceable by at least one fewer than the plurality of compute groups based on at least a period of inactivity; and

decommissioning at least one compute group of the plurality of compute groups.

5. The method of claim 1 , wherein the plurality of compute groups is allocated within a specific geographic area selected by a user.

6. The method of claim 1 , wherein the plurality of compute groups is allocated within a plurality of specific geographic region selected by a user.

7. The method of claim 1 , wherein creating the new compute group comprises:

adding the new compute group up to a predetermined maximum number of compute groups.

8. The method of claim 1 , wherein routing the one or more of the plurality of queries for the data warehouse to each of the plurality of compute groups comprises routing the one or more of the plurality of queries based on a workload of each of the plurality of compute groups.

9. The method of claim 1 , wherein each of the plurality of compute groups includes a local storage.

10. The method of claim 1 , further comprising:

dynamically managing storage resources associated with the plurality of compute groups.

11. The method of claim 1 , wherein the targeted degree of parallelism is input by a customer.

12. A non-transitory computer-readable medium for implementing a fault-tolerant data warehouse using availability zones, storing instructions which, when executed by one or more processors of a computing device, cause the one or more processors to:

allocate a plurality of compute groups each comprising a plurality of compute resources as part of a data warehouse, wherein each of the compute groups accesses data using a plurality of queries with one or more databases in one or more cloud storage resources, wherein each of the compute groups includes a plurality of processing resources;

route, by a processing device, one or more of the plurality of queries to each of the plurality of compute groups, the plurality of compute groups comprising at least two processing units in different availability zones, an availability zone comprising one or more data centers, each data center comprising redundant power, networking, and connectivity, the one or more of the plurality of queries having a common session identifier with a query previously provided to the compute group, the compute group further determined to be caching a data segment usable by the one or more of the plurality of queries, wherein in response to the one or more of the plurality of queries, each of the plurality of compute groups performs a database operation on a particular portion of a database table; and

change a total number of the plurality of compute groups using at least a comparison of a targeted degree of parallelism and a runtime computed degree of parallelism of the plurality of compute groups, wherein the runtime computed degree of parallelism is computed using a number of queries running at an input degree of parallelism and a change is at least one of creating a new compute group or deleting an existing compute group.

13. The non-transitory computer-readable medium of claim 12 , wherein the deleting the compute group comprises removing the compute group down to a predetermined minimum number of compute groups.

14. The non-transitory computer-readable medium of claim 12 , wherein the instructions of deleting the existing compute group further cause the one or more processors to:

determine whether a current workload is serviceable by fewer compute groups than the plurality of compute groups while meeting a performance metric; and

decommission at least one compute group of the plurality of compute groups in response to determining that the current workload is serviceable by fewer than the plurality of compute groups.

15. The non-transitory computer-readable medium of claim 12 , wherein the instructions of deleting the existing compute group further cause the one or more processors to:

determine whether a current workload is serviceable by at least one fewer than the plurality of compute groups based on at least a period of inactivity; and

decommission at least one compute group of the plurality of compute groups.

16. The non-transitory computer-readable medium of claim 12 , wherein the plurality of compute groups is allocated within a specific geographic area selected by a user.

17. The non-transitory computer-readable medium of claim 12 , wherein the plurality of compute groups is allocated within a plurality of specific geographic region selected by a user.

18. The non-transitory computer-readable medium of claim 12 , wherein the instructions of creating the new compute group further cause the one or more processors to:

add the new compute group up to a predetermined maximum number of compute groups.

19. The non-transitory computer-readable medium of claim 12 , wherein routing the one or more of the plurality of queries for the data warehouse to each of the plurality of compute groups comprises routing the one or more of the plurality of queries based on a workload of each of the plurality of compute groups.

20. The non-transitory computer-readable medium of claim 12 , wherein each of the plurality of compute groups includes a local storage.

21. The non-transitory computer-readable medium of claim 12 , wherein the instructions further cause the one or more processors to:

dynamically manage storage resources associated with the plurality of compute groups.

22. The non-transitory computer-readable medium of claim 12 , wherein the targeted degree of parallelism is input by a customer.

23. A system for implementing a fault-tolerant data warehouse using availability zones, comprising:

a processing device programmed to:

allocate a plurality of compute groups each comprising a plurality of compute resources as part of a data warehouse, wherein each of the compute groups accesses data using a plurality of queries with one or more databases in one or more cloud storage resources, wherein each of the compute groups includes a plurality of processing resources;

route, by the processing device, one or more of the plurality of queries to each of the plurality of compute groups, the plurality of compute groups comprising at least two processing units in different availability zones, an availability zone comprising one or more data centers, each data center comprising redundant power, networking, and connectivity, the one or more of the plurality of queries having a common session identifier with a query previously provided to the compute group, the compute group further determined to be caching a data segment usable by the one or more of the plurality of queries, wherein in response to the one or more of the plurality of queries, each of the plurality of compute groups performs a database operation on a particular portion of a database table; and

change a total number of the plurality of compute groups using at least a comparison of a targeted degree of parallelism and a runtime computed degree of parallelism of the plurality of compute groups, wherein the runtime computed degree of parallelism is computed using a number of queries running at an input degree of parallelism and a change is at least one of creating a new compute group or deleting an existing compute group.

24. The system of claim 23 , wherein the deleting the compute group comprises deleting the compute group down to a predetermined minimum number of compute groups.

25. The system of claim 23 , wherein deleting the existing compute group causes the processing device to:

determine whether a current workload is serviceable by at least fewer compute groups than the plurality of compute groups while meeting a performance metric; and

decommission at least one compute group of the plurality of compute groups in response to determining that the current workload is serviceable by fewer than the plurality of compute groups.

26. The system of claim 23 , wherein deleting the existing compute group causes the processing device to:

determine whether a current workload is serviceable by at least one fewer than the plurality of compute groups based on at least a period of inactivity; and

decommission at least one compute group of the plurality of compute groups.

27. The system of claim 23 , wherein the plurality of compute groups is allocated within a specific geographic area selected by a user.

28. The system of claim 23 , wherein the plurality of compute groups is allocated within a plurality of specific geographic region selected by a user.

29. A method for implementing a fault-tolerant data warehouse using availability zones, comprising:

allocating a plurality of compute groups each comprising a plurality of compute resources as part of a data warehouse, wherein each of the compute groups accesses data using a plurality of queries with one or more databases in one or more cloud storage resources, wherein each of the compute groups includes a local storage;

routing, by a processing device, one or more of the plurality of queries to each of the plurality of compute groups, the plurality of compute groups comprising at least two processing units in different availability zones, an availability zone comprising one or more data centers, each data center comprising redundant power, networking, and connectivity, the one or more of the plurality of queries having a common session identifier with a query previously provided to the compute group, the compute group further determined to be caching a data segment usable by the one or more of the plurality of queries, wherein in response to the one or more of the plurality of queries, each of the plurality of compute groups performs a database operation on a particular portion of a database table; and

changing storage resources independently of adding a compute group to one or more of the compute groups using an amount of free storage available for one or more of the compute groups and the plurality of queries, and a change is at least one of creating new persistent storage resources or deleting an existing persistent storage resource.

30. The system of claim 23 , wherein the targeted degree of parallelism is input by a customer.