Placement goal-based database instance dynamic consolidation

• US 8,667,020 B2
• Filed: 04/01/2011
• Issued: 03/04/2014
• Est. Priority Date: 04/01/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Zhaofu, Li (Johnson), Bin, Kuang (Bin), Aijun, Huang (Allen)
• Primary Examiner(s): Alam, Shahid
• Assistant Examiner(s): Sultana, Nargis
• Application Number:
  US13/078,318
• Publication Number:
  US 20120254435A1
• Time in Patent Office: 1,068 Days
• Field of Search: 707/790, 709/226
• US Class Current: 707/790
• CPC Class Codes: G06F 16/20 of structured data, e.g. relational data
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers

Abstract

Models and tools for facilitating database instance consolidation are described. One approach obtains standardized database instance performance data, receives instance isolation criteria which are applied to divide instances into sets, determines resource usage for each set, and assigns instance sets to target machines based on resource usage and capacity. Each instance set is associated with a virtual machine in some cases. A security isolation criterion indicates that specified instances shall not co-reside on a target machine; an organizational criterion indicates that specified instances shall be located together in a single data center; a compatibility criterion indicates that instances shall not reside in the same virtual machine, the same physical server, or the same server pool. Users iteratively reach an acceptable consolidation of instances at one or more data centers, such as an enhanced assignment of instances that is more efficient than a linear additive solution.

Claims

1. A computer-readable storage medium configured with data and with instructions that when executed by at least one processor causes the processor(s) to perform a process for facilitating database instance consolidation, the process comprising the steps of:

obtaining standardized database instance performance data for database instances running on source machines;

receiving at least one database instance isolation criterion, wherein at least one isolation criterion includes at least one of the following;

(a) a security factor indicating that specified instances shall not co-reside on a target machine, (b) an organizational factor indicating that specified instances shall be located together in a single data center, (c) a compatibility factor indicating that specified instances shall not reside in the same virtual machine, (d) a compatibility factor indicating that specified instances shall not reside in the same physical server, or (e) a compatibility factor indicating that specified instances shall not reside in the same server pool;

applying at least one received database isolation criterion to automatically divide database instances into instance sets;

determining a resource usage for each instance set; and

automatically assigning instance sets to target machines based on instance set resource usage and target machine resource capacity, and utilizing at least one of the following;

dynamic programming, combinatorial optimization.

2. The configured medium of claim 1, wherein at least one of the following is received as a database instance isolation criterion:

a security isolation criterion indicating that specified instances shall not co-reside on a target machine;

an organizational isolation criterion indicating that specified instances shall be located together in a single data center.

3. The configured medium of claim 1, wherein determining a resource usage for each instance set comprises rounding up to a multiple of a predetermined utilization unit.

4. The configured medium of claim 1, wherein assigning instance sets to target machines comprises sorting target machines by resource capacity.

5. The configured medium of claim 1, wherein the determining step utilizes standardized performance data.

6. The configured medium of claim 1, wherein the process further comprises associating each instance set with a virtual machine, and assigning instance sets includes assigning virtual machines to target machines.

7. The configured medium of claim 6, wherein determining a resource usage for each instance set comprises finding a virtual machine resource usage for each virtual machine.

8. The configured medium of claim 1, wherein assigning instance sets to target machines comprises solving a 0-1 knapsack problem.

9. A process for placement goal-based database instance consolidation, the process comprising the steps of:

gathering database instance performance data for database instances running on source machines;

standardizing gathered performance data;

entering into a user interface of a consolidation tool at least one database instance placement goal in the form of at least one isolation criterion, wherein at least one isolation criterion includes at least one of the following;

(a) a security factor indicating that specified instances shall not co-reside on a target machine, (b) an organizational factor indicating that specified instances shall be located together in a single data center, (c) a compatibility factor indicating that specified instances shall not reside in the same virtual machine, (d) a compatibility factor indicating that specified instances shall not reside in the same physical server, or (e) a compatibility factor indicating that specified instances shall not reside in the same server pool;

identifying target machine characteristics to the consolidation tool;

commanding the consolidation tool to apply one or more database isolation criteria to the standardized performance data and the target machine characteristics; and

getting from the consolidation tool as an output an enhanced assignment of instances to target machines which is more resource efficient than an additive linear assignment.

10. The process of claim 9, wherein standardizing gathered performance data includes at least one of the following:

setting a Needed Compute Unit to a Normalized CPU Capacity times a Processor Time Percentage;

setting Needed Compute Unit to a Normalized CPU Capacity times a Processor Time Percentage times a multiplier which exceeds one;

setting a Virtual Machine Compute Unit to a Target Machine Compute Unit divided by a Target Machine Number of Processors and times a Virtual Machine Number of Processors, when Virtual Machine Number of Processors does not exceed Target Machine Number of Processors;

setting a Virtual Machine Compute Unit to a Target Machine Compute Unit when a Virtual Machine Number of Processors exceeds a Target Machine Number of Processors;

setting a Virtual Machine Memory Usage based on a Virtual Machine Startup Memory Allocation and a Virtual Machine Maximum Memory Allocation.

11. The process of claim 9, wherein the entering step comprises entering at least one isolation criterion into a spreadsheet program, and the spreadsheet program serves as at least a portion of the user interface of the consolidation tool.

12. The process of claim 9, wherein the getting step gets an assignment of instances for at least one of the following situations:

source machines are treated as being located in multiple data centers and target machines are treated as being located in multiple data centers;

source machines are treated as being located in multiple data centers and target machines are treated as being located in a single data center;

source machines are treated as being located in a single data center and target machines are treated as being located in multiple data centers;

source machines are treated as being located in a single data center and target machines are treated as being located in a single data center.

13. The process of claim 9, wherein the entering step comprises entering at least three isolation criteria.

14. The process of claim 9, wherein the entering step comprises entering a security isolation criterion which indicates that specified instances shall not co-reside on a target machine.

15. The process of claim 9, wherein the entering step comprises entering an organizational isolation criterion which indicates that specified instances shall be located together in a single data center.

16. A computer system configured to assist database instance consolidation, the system comprising:

a logical processor;

a memory in operable communication with the logical processor;

at least one database instance isolation criterion residing in the memory, wherein at least one isolation criterion includes at least one of the following;

(a) a security factor indicating that specified instances shall not co-reside on a target machine, (b) an organizational factor indicating that specified instances shall be located together in a single data center, (c) a compatibility factor indicating that specified instances shall not reside in the same virtual machine, (d) a compatibility factor indicating that specified instances shall not reside in the same physical server, or (e) a compatibility factor indicating that specified instances shall not reside in the same server pool; and

database instance consolidation code residing in the memory and capable of automatically applying the database isolation criterion to divide database instances into instance sets and capable of automatically assigning instance sets to target machines by dynamic programming based on instance set resource usage and target machine resource capacity.

17. The system of claim 16, wherein the system is further characterized by standardized database instance performance data residing in the memory, and the database instance consolidation code is also capable of determining a resource usage for each instance set based on the performance data.

18. The system of claim 16, wherein the database instance consolidation code comprises a spreadsheet program which is programmed to automatically assign instance sets to target machines based on instance set resource usage and target machine resource capacity.

19. The system of claim 16, wherein the database instance consolidation code is programmed to assign instance sets to target machines based at least in part on a solution of a 0-1 knapsack problem arising at least in part from the instance sets.

20. The system of claim 16, further characterized in that an enhanced assignment of instances to target machines resides in the memory, namely, an assignment which is more resource efficient than an additive linear assignment.