Dynamically varying a raid cache policy in order to optimize throughput

• US 20050076115A1
• Filed: 09/24/2003
• Published: 04/07/2005
• Est. Priority Date: 09/24/2003
• Status: Active Grant

• Current Assignee: Dell Products LP (Dell Technologies Inc.)
• Original Assignee: Dell Products LP (Dell Technologies Inc.)
• Inventors: Andrews, Carlton A., Vrhel, Thomas Jr.
• Granted Patent: US 7,487,235 B2
• Time in Patent Office: Days
• US Class Current: 709/224
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/1025 Dynamic adaptation of the criteria on which the server selection is based

Abstract

In an information handling system, an application operating on an elements of a the information handling system having one or more RAID systems is used to monitor the number of client devices and other operational aspects of the system. The application is constructed and arranged to determine and implement the best cache policy for the RAID systems in order to maximize the desired operational characteristics of the information handling system. The application can operate on a single server or RAID system device, or the application can operate on a separate device to control the cache policies of one or more servers.

Claims

1. An information handling system comprising:

a server, said sever having one or more RAID systems, said RAID systems capable of implementing two or more cache policies;

a network operative with said server, said network connecting one or more clients to said server, said clients constructed and arranged to communicate with said server thereby placing a load on said server; and

a load monitor operative with said one or more RAID systems, said load monitor constructed and arranged to monitor said load, said load monitor further constructed and arranged to select one or more cache policies of said one or more RAID systems that optimize a performance characteristic of said information handling system;

wherein said load manager monitors said load and implements a cache policy that optimizes a characteristic of said information handling system.

2. The information handling system according to claim 1, wherein said load monitor employs a template to select one or more cache policies.

3. The information handling system according to claim 1, wherein said load monitor employs an algorithm to select said one or more cache policies.

4. The information handling system according to claim 1, wherein said load monitor employs a template and an algorithm to select said one or more cache policies.

5. The information handling system according to claim 1, wherein said RAID system has a read cache.

6. The information handling system according to claim 5, wherein said read cache has a no-ahead policy.

7. The information handling system according to claim 5, wherein said read cache has an adaptive policy.

8. The information handling system according to claim 5, wherein said read cache has a read-ahead policy.

9. The information handling system according to claim 1, wherein said RAID system has a write cache.

10. The information handling system according to claim 9, wherein said write cache has back policy.

11. The information handling system according to claim 9, wherein said write cache has through policy.

12. The information handling system according to claim 1, wherein said RAID system has an I/O.

13. The information handling system according to claim 12, wherein said I/O has a cached policy.

14. The information handling system according to claim 12, wherein said I/O has a direct policy.

15. The information handling system according to claim 1, wherein one of said cache policies is a no-ahead policy.

16. The information handling system according to claim 1, wherein one of said cache policies is an adaptive policy.

17. The information handling system according to claim 1, wherein one of said cache policies is a read-ahead policy.

18. The information handling system according to claim 1, wherein one of said cache policies is back policy.

19. The information handling system according to claim 1, wherein one of said cache policy is a through policy.

20. The information handling system according to claim 1, wherein said cache policy is a cached policy.

21. The information handling system according to claim 1, wherein said load monitor is a load balancer.

22. The information handling system according to claim 1, wherein said load monitor is a router.

23. The information handling system according to claim 1, wherein said load monitor is a server.

24. The information handling system according to claim 1, wherein said load monitor is a cluster master.

25. An information handling system comprising:

at least one server, said sever having one or more RAID systems, said RAID systems capable of implementing two or more cache policies;

a network operative with said server, said network connecting one or more clients to said server, said clients constructed and arranged to communicate with said server thereby placing a load on said server;

a load balancer, said load balancer constructed and arranged to allocate said load among said one or more servers; and

a load monitor operative on said load balancer, said load monitor constructed and arranged to monitor said load, said load monitor further constructed and arranged to select a cache policy of said one or more RAID systems on said one or more servers that optimize a performance characteristic of said information handling system;

wherein said load manager monitors said load and implements a cache policy that optimizes a characteristic of said information handling system.

26. The information handling system according to claim 25, wherein said load monitor is a load balancer.

27. The information handling system according to claim 25, wherein said load monitor is a router.

28. The information handling system according to claim 25, wherein said load monitor is a server.

29. The information handling system according to claim 25, wherein said load monitor is a cluster master.

30. A method for changing cache policy for a RAID system on an information handling system comprising:

a) reading a set of templates;

b) determining the load of the network;

c) indexing said templates using said load to determine a cache setting; and

d) applying said cache settings to said RAID system.

31. The method according to claim 30, further comprising:

e) invoking a delay.

32. The method according to claim 30, wherein said steps a) through d) are performed continuously on said information handling system.

33. The method according to claim 30, wherein said steps a) through d) are performed continuously in real-time on said information handling system.

34. The method according to claim 30, wherein said steps a) through d) are performed continuously in near-real-time on said information handling system.

35. The method according to claim 30, wherein said step of determining employs a template to select said cache setting.

36. The method according to claim 30, wherein said step of determining employs an algorithm to select said cache setting.

37. The method according to claim 30, wherein said step of determining employs a template and an algorithm to select said cache setting.

38. The method according to claim 30, wherein said cache setting is made up of two or more policies.

39. The method according to claim 38, wherein said RAID system has a read cache.

40. The method according to claim 39, wherein said read cache has a no-ahead policy.

41. The method according to claim 39, wherein said read cache has an adaptive policy.

42. The method according to claim 39, wherein said read cache has a read-ahead policy.

43. The method according to claim 38, wherein said RAID system has a write cache.

44. The method according to claim 43, wherein said write cache has back policy.

45. The method according to claim 43, wherein said write cache has through policy.

46. The method according to claim 38, wherein said RAID system has an I/O.

47. The method according to claim 46, wherein said I/O has a cached policy.

48. The method according to claim 46, wherein said I/O has a direct policy.

49. The method according to claim 30, wherein said cache setting includes a no-ahead policy.

50. The method according to claim 30, wherein said cache setting includes an adaptive policy.

51. The method according to claim 30, wherein said cache setting includes a read-ahead policy.

52. The method according to claim 30, wherein said cache setting includes a back policy.

53. The method according to claim 30, wherein said cache setting includes a through policy.

54. The method according to claim 30, wherein said cache setting includes a cached policy.

55. The method according to claim 30, wherein said load monitor is a load balancer.

56. The method according to claim 30, wherein said load monitor is a router.

57. The method according to claim 30, wherein said load monitor is a server.

58. The method according to claim 30, wherein said load monitor is a cluster master.

59. An information handling system comprising:

client means for generating a load;

server means for servicing said load, said sever having one or more RAID systems, said RAID systems capable of implementing two or more cache policies;

network means operative with said server means and said client means for transmitting said load; and

load monitor means for monitoring said load, said load monitor means operative with said one or more RAID systems, said load monitor constructed and arranged to select a cache policy of said one or more RAID systems that optimize a performance characteristic of said information handling system;

wherein said load manager monitors said load and implements a cache policy that optimizes a characteristic of said information handling system.