Facilitating tiered service model-based fair allocation of resources for application servers in multi-tenant environments

• US 10,169,090 B2
• Filed: 02/27/2015
• Issued: 01/01/2019
• Est. Priority Date: 09/12/2012
• Status: Active Grant

• Current Assignee: Salesforce.com, Inc.
• Original Assignee: Salesforce.com, Inc.
• Inventors: Wang, Xiaodan
• Primary Examiner(s): Bullock, Jr., Lewis Alexander
• Assistant Examiner(s): Headly, Melissa A
• Application Number:
  US14/634,289
• Publication Number:
  US 20150178135A1
• Time in Patent Office: 1,404 Days
• Field of Search: None
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals
• CPC Class Codes: G06F 9/5088 involving task migration
• CPC Class Codes: H04L 41/5051 Service on demand, e.g. definition and deployment of services in real time
• CPC Class Codes: H04L 41/5096 wherein the managed service relates to distributed or central networked applications
• CPC Class Codes: H04L 45/021 Ensuring consistency of routing table updates, e.g. by using epoch numbers
• CPC Class Codes: H04L 45/742 Route cache; Operation thereof

Abstract

In accordance with embodiments, there are provided mechanisms and methods for facilitating tiered service model-based fair allocation of resources for application servers in multi-tenant environments. In one embodiment and by way of example, a method includes collecting, by and incorporating into the database system, data relating to job types associated with one or more tenants of a plurality of tenants within a multi-tenant database system, computing, based on the data, an actual resource use and an expected resource allocation associated with each job type, and assigning classifications to the job types based on their corresponding actual resource use and the expected resource allocation. The method may further include routing the job types between tiers based on the assigned classifications, where the routing includes at least one of promoting, demoting, and maintaining one or more tiers for the job types.

Claims

1. A method comprising:

collecting, by a resource-management server computing device of a database system, data relating to job types associated with multiple tenants within a multi-tenant environment;

based on the data, computing an actual resource usages and expected resource allocations of the job types and actual resource usages and expected resource allocations of the tenants;

assigning the job types to service tiers based on the actual resource usages and the expected resource allocations associated with the job types, wherein each job type is at least one of a high-tiered job type or a low-tiered job type;

assigning the tenants to the service tiers based on the actual resource usages and the expected resource allocations associated with the tenants, wherein each tenant is classified as a high-tiered tenant type or a low-tiered tenant type; and

real-time reassigning and executing of the job types to one or more of the service tiers while ensuring that resources are distributed between the job types and the tenants such that actual resource usage does not exceed expected resource allocation for each job type and each tenant.

2. The method of claim 1, wherein the data comprises statistics relating to at least one of thread time, queueing time, processing jobs, and failed jobs relating to the job types associated with the tenants.

3. The method of claim 1, further comprising determining a resource-consumption factor for a job type or a tenant based on an actual resource usage and an expected resource allocation of the job type or the tenant, respectively.

4. The method of claim 1, wherein the job types and the tenants are assigned to service tiers based on one or more resource-consumption factors indicating consumption of the resources by at least one of the job types and the tenants, wherein the resource-consumption factors include at least one of 1) victim, when the actual resource usages of the job types or the tenants are less than the expected resource allocation, 2) offender, when the actual resource usages of the job types or the tenants are greater than the expected resource allocation, and 3) fair, when the actual resource usages of the job types or the tenants are equal to the expected resource allocation.

5. The method of claim 4, wherein each job type is classified as at least one type of multi-tiered job types including one or more of the high-tiered job type and the low-tiered job type.

6. The method of claim 4, wherein each tenant is classified as at least one type of multi-tiered tenant types including one or more of the high-tiered tenant type and tenant type.

7. The method of claim 1, wherein the job types are requested through job requests received from the tenants and submitted by users representing the tenants, wherein the job requests are placed by the users via one or more computing devices over one or more networks, wherein the one or more network include a Cloud-based network or the Internet.

8. A system comprising:

a processor and a memory to execute instructions at the system; and

a mechanism, wherein the processor to facilitate the mechanism to;

collect data relating to job types associated with multiple tenants within a multitenant database environment;

compute, based on the collected data, actual resource usages and expected resource allocations of the job types and actual resource usages and expected resource allocations of the tenants;

assign the job types to service tiers based on the actual resource usages and the expected resource allocations associated with the job types, wherein each job type is at least one of a high-tiered job type or a low tiered job type;

assign the tenants to the service tiers based on the actual resource usages and the expected resource allocations associated with the tenants, wherein each tenant is classified as a high-tiered tenant or a low-tiered tenant type; and

in real time, reassigning and executing of the job types to one or more of the service tiers while ensuring that resources are distributed between the job types and the tenants such that the actual resource usages does not exceed the expected resource allocation for each job type and each tenant.

9. The system of claim 8, wherein the data comprises statistics relating to at least one of thread time, queueing time, processing jobs, and failed jobs relating to the job types associated with the tenants.

10. The system of claim 8, wherein the mechanism is further to determine a resource-consumption factor for a job type or a tenant based on an actual resource usage and an expected resource allocation of the job type or the tenant, respectively.

11. The system of claim 8, wherein the job types and the tenants are assigned to service tiers based on one or more resource-consumption factors indicating consumption of the resources by at least one of the job types and the tenants, wherein the resource-consumption factors include at least one of 1) victim, when the actual resource usages of the job types or the tenants are less than the expected resource allocation, 2) offender, when the actual resource usages of the job types or the tenants are greater than the expected resource allocation, and fair, when the actual resource usages of the job types or the tenants are equal to the expected resource allocation.

12. The system of claim 11, wherein each job type is classified as at least one type of multi-tiered job types including one or more of the high-tiered job type and the low-tiered job type.

13. The system of claim 11, wherein each tenant is classified as at least one type of multi-tiered tenant types including one or more of the high-tiered tenant type and the low-tiered tenant type.

14. The system of claim 8, wherein the job types are requested through job requests received from the tenants and submitted by users representing the tenants, wherein the job requests are placed by the users via one or more computing devices over one or more networks, wherein the one or more networks include a Cloud-based network or the Internet.

15. A non-transitory machine-readable medium comprising a plurality of instructions which, when executed by a processing device, cause the processing device to perform operations comprising:

collecting data relating to job types associated with multiple tenants within a multi-tenant database environment;

computing, based on the collected data, an actual resource usages and expected resource allocations of the job types and actual resource usages and expected resource allocations of the tenants;

assigning the job types to service tiers based on the actual resource usages and the expected resource allocations associated with the job types;

assigning the tenants to the service tiers based on the actual resource usages and the expected resource allocations associated with the tenants, wherein each tenant is classified as a high-tiered tenant type or a low-tiered tenant type; and

real-time reassigning and executing of the job types to one or more of the service tiers while ensuring that resources are distributed between the job types and the tenants such that actual resource usage does not exceed expected resource allocation for each job type and each tenant.

16. The machine-readable medium of claim 15, wherein the data comprises statistics relating to at least one of thread time, queueing time, processing jobs, and failed jobs relating to the job types associated with the tenants.

17. The machine-readable medium of claim 15, wherein the operations comprise determining a resource-consumption factor a job type or a tenant based on an actual resource usage and an expected resource allocation of the job type or the tenant, respectively.

18. The machine-readable medium of claim 15, wherein the job types and the tenants are assigned to service tiers based on one or more resource-consumption factors indicating consumption of the resources by at least one of the job types and the tenants, wherein the resource-consumption factors include at least one of 1) victim, when the actual resource usages of the job types or the tenants are less than the expected resource allocation, 2 offender, when the actual resource usages of the job types or the tenants are greater than the expected resource allocation, and 2 fair, when the actual resource usages of the job types or the tenants are equal to the expected resource allocation.

19. The machine-readable medium of claim 18, wherein each job type is classified as at least one type of multi-tiered job types including one or more of the high-tiered job type and the low-tiered job type.

20. The machine-readable medium of claim 18, wherein each tenant is classified as at least one type of multi-tiered tenant types including one or more of the high-tiered tenant type and the low-tiered tenant type.

21. The machine-readable medium of claim 15, wherein the job types are requested through job requests received from the tenants and submitted by users representing the tenants, wherein the job requests are placed by the users via one or more computing devices over one or more networks, wherein the one or more networks include a Cloud-based network or the Internet.