WORKLOAD-DRIVEN TECHNIQUES FOR PROVIDING BIASED SERVICE LEVEL GUARANTEES

• US 20160224380A1
• Filed: 01/30/2015
• Published: 08/04/2016
• Est. Priority Date: 01/30/2015
• Status: Active Grant

• Current Assignee: Salesforce.com, Inc.
• Original Assignee: Salesforce.com, Inc.
• Inventors: Wang, Xiaodan
• Granted Patent: US 9,766,960 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 2209/548 Queue
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

Techniques and architectures for workload management. A stream of messages is received with servers coupled to provide access to shared system resources. The servers are organized as at least two groups with a first group providing message queuing services and a second group providing message queuing services under first conditions and modified message queuing services under second conditions. Message queuing services are provided with the first group by treating all jobs as equal, processing jobs asynchronously and processing jobs in the background to defer to foreground, non-message queuing services, jobs. Message queuing services are provided with the second group by treating all jobs as equal, processing jobs asynchronously and processing jobs in the background to defer to foreground, non-message queuing services, jobs under the first conditions. Modified message queuing services are provided with the second group by shortest-slack-first scheduling under the second conditions.

Claims

1. A system comprising:

one or more shared system resources;

a plurality of servers coupled to provide access to the one or more shared system resources, the plurality of servers organized as at least two groups with a first group providing message queuing services and a second group providing message queuing services under first conditions and modified message queuing services under second conditions;

wherein the first group provides message queuing services by treating all jobs as equal, processing jobs asynchronously and processing jobs in the background to defer to foreground, non-message queuing services, jobs;

wherein the second group provides message queuing services by treating all jobs as equal, processing jobs asynchronously and processing jobs in the background to defer to foreground, non-message queuing services, jobs under the first conditions; and

wherein the second group provides modified message queuing services by scheduling message queue processing based on message type and associated service level parameters.

2. The system of claim 1 wherein the second group further provides modified message queuing services by utilizing shortest-slack-first scheduling under the second conditions.

3. The system of claim 1 wherein the second group further provides modified message queuing services by utilizing selective load shedding under the second conditions.

4. The system of claim 1 wherein the second group further provides modified message queuing services by utilizing dynamic message reordering under the second conditions.

5. The system of claim 1 wherein the second group further provides modified message queuing services by utilizing segmented clustering under the second conditions.

6. The system of claim 1 wherein jobs are classified by a delay threshold (DT), which is a tolerance to job processing delays.

7. The system of claim 1 wherein each message has an associated slack, which corresponds to a deadline for completion of the corresponding message.

8. The system of claim 1 wherein, for messages received by the plurality of servers, dynamic features of the messages that are captured comprise a number of messages for a specific type in a queue, an average per message run time, and an arrival rate of new messages of a specific type.

9. The system of claim 8 wherein the dynamic features of the messages are utilized to determine whether there exists sufficient messaging queue capacity to meet service level guarantees.

10. The system of claim 1 wherein the first conditions comprise meeting service level agreements and the second conditions comprise not meeting service level agreements.

11. The system of claim 1 wherein the first conditions comprise prediction that service level agreements will be met and second conditions comprise prediction that service level agreements will not be met.

12. The system of claim 1 wherein the one or more shared system resources comprises at least an on demand services environment system.

13. The system of claim 12 wherein the on demand services environment comprises a multitenant database environment.

15. The non-transitory computer-readable medium of claim 1 wherein the second group further provides modified message queuing services by utilizing shortest-slack-first scheduling under the second conditions.

14. A non-transitory computer-readable medium having stored there on instructions that, when executed by one or more processors, cause the one or more processors to:

receive a stream of messages with a plurality of servers coupled to provide access to the one or more shared system resources, the plurality of servers organized as at least two groups with a first group providing message queuing services and a second group providing message queuing services under first conditions and modified message queuing services under second conditions;

provide message queuing services with the first group by treating all jobs as equal, processing jobs asynchronously and processing jobs in the background to defer to foreground, non-message queuing services, jobs;

provide message queuing services with the second group by treating all jobs as equal, processing jobs asynchronously and processing jobs in the background to defer to foreground, non-message queuing services, jobs under the first conditions; and

provide modified message queuing services with the second group by scheduling message queue processing based on message type and associated service level parameters.

16. The non-transitory computer-readable medium of claim 14 wherein the second group further provides modified message queuing services by utilizing selective load shedding under the second conditions.

17. The non-transitory computer-readable medium of claim 14 wherein the second group further provides modified message queuing services by utilizing dynamic message reordering under the second conditions.

18. The non-transitory computer-readable medium of claim 14 wherein the second group further provides modified message queuing services by utilizing segmented clustering under the second conditions.

19. The non-transitory computer-readable medium of claim 14 wherein jobs are classified by a delay threshold (DT), which is a tolerance to job processing delays.

20. The non-transitory computer-readable medium of claim 14 wherein each message has an associated slack, which corresponds to a deadline for completion of the corresponding message.

21. The non-transitory computer-readable medium of claim 14 wherein, for messages received by the plurality of servers, dynamic features of the messages that are captured comprise a number of messages for a specific type in a queue, an average per message run time, and an arrival rate of new messages of a specific type.

22. The non-transitory computer-readable medium of claim 21 wherein the dynamic features of the messages are utilized to determine whether there exists sufficient messaging queue capacity to meet service level guarantees.

23. The non-transitory computer-readable medium of claim 14 wherein the first conditions comprise meeting service level agreements and the second conditions comprise not meeting service level agreements.

24. The non-transitory computer-readable medium of claim 14 wherein the first conditions comprise prediction that service level agreements will be met and second conditions comprise prediction that service level agreements will not be met.

25. The non-transitory computer-readable medium of claim 14 wherein the one or more shared system resources comprises at least an on demand services environment system.

26. The non-transitory computer-readable medium of claim 25 wherein the on demand services environment comprises a multitenant database environment.

27. A method comprising:

receiving a stream of messages with a plurality of servers coupled to provide access to the one or more shared system resources, the plurality of servers organized as at least two groups with a first group providing message queuing services and a second group providing message queuing services under first conditions and modified message queuing services under second conditions;

providing message queuing services with the first group by treating all jobs as equal, processing jobs asynchronously and processing jobs in the background to defer to foreground, non-message queuing services, jobs;

providing message queuing services with the second group by treating all jobs as equal, processing jobs asynchronously and processing jobs in the background to defer to foreground, non-message queuing services, jobs under the first conditions; and

providing modified message queuing services with the second group by scheduling message queue processing based on message type and associated service level parameters.

28. The method of claim 27 wherein the second group further provides modified message queuing services by utilizing shortest-slack-first scheduling under the second conditions.

29. The method of claim 27 wherein the second group further provides modified message queuing services by utilizing selective load shedding under the second conditions.

30. The method of claim 27 wherein the second group further provides modified message queuing services by utilizing dynamic message reordering under the second conditions.

31. The method of claim 27 wherein the second group further provides modified message queuing services by utilizing segmented clustering under the second conditions.

32. The method of claim 27 wherein jobs are classified by a delay threshold (DT), which is a tolerance to job processing delays.

33. The method of claim 27 wherein each message has an associated slack, which corresponds to a deadline for completion of the corresponding message.

34. The method of claim 27 wherein, for messages received by the plurality of servers, dynamic features of the messages that are captured comprise a number of messages for a specific type in a queue, an average per message run time, and an arrival rate of new messages of a specific type.

35. The method of claim 34 wherein the dynamic features of the messages are utilized to determine whether there exists sufficient messaging queue capacity to meet service level guarantees.

36. The method of claim 27 wherein the first conditions comprise meeting service level agreements and the second conditions comprise not meeting service level agreements.

37. The method of claim 27 wherein the first conditions comprise prediction that service level agreements will be met and second conditions comprise prediction that service level agreements will not be met.