Managing data center resources to achieve a quality of service

• US 20140095691A1
• Filed: 09/28/2012
• Published: 04/03/2014
• Est. Priority Date: 09/28/2012
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Ganguli, Mrittika, Moses, Jaideep, Herdrich, Andrew J., I., Muthuvel M., Narayan, Ananth S., Khanna, Rahul
• Granted Patent: US 10,554,505 B2
• Time in Patent Office: Days
• US Class Current: 709/224
• CPC Class Codes: H04L 41/145 involving simulating, designing, planning or modelling of a network
• CPC Class Codes: H04L 41/50 Network service management, e.g. ensuring proper service fulfilment according to agreements
• CPC Class Codes: H04L 41/5009 Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF]
• CPC Class Codes: H04L 41/5025 by proactively reacting to service quality change, e.g. by reconfiguration after service quality degradation or upgrade

Abstract

In accordance with some embodiments, a cloud service provider may operate a data center in a way that dynamically reallocates resources across nodes within the data center based on both utilization and service level agreements. In other words, the allocation of resources may be adjusted dynamically based on current conditions. The current conditions in the data center may be a function of the nature of all the current workloads. Instead of simply managing the workloads in a way to increase overall execution efficiency, the data center instead may manage the workload to achieve quality of service requirements for particular workloads according to service level agreements.

Claims

1. One or more computer readable media storing instructions executed by a data center, including a plurality of processors, to perform a sequence:

analyzing data from a performance monitor from one of said processors;

developing a performance metric based on said data; and

using the performance metric and a quality of service value to dynamically reallocate resources within the data center.

2. The media of claim 1, said sequence including receiving performance monitor data from at least two cores, each core providing different types of performance monitor data.

3. The media of claim 1, said sequence including analyzing performance monitor data about instructions per cycle.

4. The media of claim 1, said sequence receiving performance monitor data including cache misses.

5. The media of claim 1, said sequence including receiving performance monitor data about memory bandwidth.

6. The media of claim 1, said sequence including where analyzing data includes analyzing data from a performance monitor to determine instructions per cycle and cache misses.

7. The media of claim 1, said sequence including using the performance metric and quality of service value to reallocate lower level caches within a data center.

8. The media of claim 1, said sequence including reallocating cache space from one node in the data center to another node in the data center.

9. The media of claim 1, said sequence including dynamically reallocating data on a periodic basis.

10. A method comprising:

analyzing data from a performance monitor;

developing a performance metric based on said data; and

using the performance metric and a quality of service value to dynamically reallocate resources within a data center.

11. The method of claim 10 including receiving performance monitor data from at least two cores, each core providing different types of performance monitor data.

12. The method of claim 10 including analyzing performance monitor data about instructions per cycle.

13. The method of claim 10 including receiving measurements of cache occupancy.

14. The method of claim 10 including receiving performance monitor data about memory bandwidth.

15. The method of claim 10 where analyzing data includes analyzing data from a performance monitor to predict instructions per cycle and cache misses.

16. The method of claim 10 including using the performance metric and quality of service value to set a policy to reallocate lower level caches within a data center to different virtual machines.

17. The method of claim 16 including reallocating cache space from one node in the data center to another node in the data center.

18. The method of claim 10 including dynamically reallocating on a periodic basis.

19. The method of claim 18, wherein dynamically reallocating resources is dependent on quality of service factors and de-emphasizes traditional efficiency allocations.

20. A server comprising:

a performance monitor;

a processor to analyze data from the performance monitor, develop a performance metric based on said data, use the performance metric and a quality of service value to reallocate resources within a data center; and

a storage coupled to said processor.

21. The server of claim 20, said processor to analyze performance monitor data about instructions per cycle.

22. The server of claim 20, said processor to receive performance monitor data including cache misses.

23. The server of claim 20, said processor to receive performance monitor data about memory bandwidth.

24. The server of claim 20, said processor to analyze data includes analyzing data from a performance monitor to determine instructions per cycle and cache misses.

25. The server of claim 20, said processor to use the performance metric and quality of service value to reallocate lower level caches within a data center.

26. The server of claim 20, said processor to reallocate cache space from one node in the data center to another node in the data center.

27. The server of claim 20, said processor to dynamically reallocate data on a periodic basis.

28. A cloud management environment comprising:

a first server including a first performance monitor to produce a first data type;

a second server coupled to said first server including a second performance monitor to produce a second data type different from said first data type; and

a processor to analyze data from each performance monitor, develop a performance metric based on said data, and use the performance metric and a quality of service value to reallocate resources between said servers.

29. The environment of claim 28, said first and second performance monitors to monitor at least one of instructions per cycle, cache misses, or memory bandwidth.

30. The environment of claim 28, said processor to use the performance metric and quality of service value to reallocate lower level caches within the environment.

31. The environment of claim 28, said processor to dynamically reallocate data on a periodic basis.