Cost optimization in dynamic workload capping

• US 9,680,657 B2
• Filed: 08/31/2015
• Issued: 06/13/2017
• Est. Priority Date: 08/31/2015
• Status: Active Grant

• Current Assignee: BMC Software Incorporated (KKR & Co., Inc.)
• Original Assignee: BMC Software Incorporated (KKR & Co., Inc.)
• Inventors: Spicer, Paul Charles, Degrange, Steven, Rama, Hemanth, Gangavarapu, Sridhar, Perini, Robert, Williams, Edward
• Primary Examiner(s): Chen, George
• Application Number:
  US14/840,708
• Publication Number:
  US 20170063561A1
• Time in Patent Office: 652 Days
• Field of Search: 705 11, 705 735, 705400-412, 718101-105
• CPC Class Codes: G06F 2209/504 Resource capping
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455
• CPC Class Codes: G06Q 30/04 Billing or invoicing
• CPC Class Codes: H04L 12/1435 volume-based
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A mainframe computing system includes a central processor complex, a plurality of billing entities, each billing entity having a respective capacity limit, and a workload manager that schedules work requested by the plurality of billing entities on the central processor complex and tracks, by billing entity, a rolling average of millions of service units. The mainframe also includes a dynamic capping policy that identifies a maximum MSU limit, a maximum cost limit, a subset of the plurality of billing entities, and, for each billing entity in the subset, information from which to determine a MSU entitlement value and cost entitlement value. The mainframe also includes a dynamic capping master that adjusts the respective capacity limits of the subset of the plurality of billing entities at scheduled intervals based on the dynamic capping policy to favor billing entities having high-importance workload within the maximum MSU limit and maximum cost limit.

Claims

1. A mainframe computing system comprising:

a central processor complex;

a plurality of billing entities, a billing entity being a logical partition of the mainframe computing system or a group of logical partitions, each billing entity having a respective capacity limit;

a workload manager that schedules work requested by the plurality of billing entities on the central processor complex and tracks, by billing entity, a rolling average of millions of service units (MSUs);

a dynamic capping policy for the central processor complex stored in non-transitory memory, the dynamic capping policy identifying;

a maximum MSU limit,

a maximum cost limit,

a subset of the plurality of billing entities, and

for each identified billing entity, information from which to determine an MSU entitlement value and a cost entitlement value; and

a dynamic capping master that monitors and adjusts the respective capacity limits of the subset of the plurality of billing entities, wherein the workload manager schedules work within the respective capacity limits so that the central processor complex executes the work without exceeding the maximum cost limit and the MSU limit, the monitoring and adjusting being based on the dynamic capping policy and bounded by cost entitlement unused from billing entities having respective rolling averages that are less than respective MSU entitlement values.

2. The mainframe computing system of claim 1, wherein adjusting the respective capacity limits includes:

determining a pool of service units that can be shared, the pool of service units that can be shared representing MSUs from billing entities that have respective rolling averages that are less than respective MSU entitlement values;

determining a cost pool for the pool of service units, the cost pool representing cost entitlement unused from billing entities that have respective rolling averages that are less than respective MSU entitlement values;

determining a favored billing entity of the subset of the plurality of billing entities that has a rolling average greater than its respective MSU entitlement value and lowest low-importance workload percentage; and

adjusting the capacity limit for the favored billing entity above the MSU entitlement value using borrowed MSUs, the borrowed MSUs being bounded by the pool of service units and the cost pool.

3. The mainframe computing system of claim 2, wherein the capacity limit for a billing entity that has a rolling average less than the MSU entitlement is set to the rolling average plus a buffer value and the difference between the capacity limit and the MSU entitlement is added to the pool of service units and the product of an MSU cost and the difference between the capacity limit and the MSU entitlement is added to the cost pool.

4. The mainframe computing system of claim 2, wherein bounding the borrowed MSUs includes:

determining a borrowed cost for the borrowed MSUs, the borrowed cost being equal to the product of the borrowed MSUs and an MSU cost for the favored billing entity;

determining whether the borrowed cost is greater than the cost pool;

reducing the cost pool by the borrowed cost when the borrowed cost is not greater than the cost pool; and

reducing the borrowed MSUs when the borrowed cost is greater than the cost pool.

5. The mainframe computing system of claim 1, wherein the cost entitlement is calculated from an MSU cost for the identified billing entity and the MSU entitlement value.

6. The mainframe computing system of claim 5, wherein the MSU cost represents a relative value.

7. The mainframe computing system of claim 5, wherein the MSU cost represents a value calculated based on a prior bill for the mainframe computing system.

8. A method of monitoring and adjusting capacity limits for billing entities identified in a dynamic capping policy on a mainframe, the dynamic capping policy including, for each billing entity identified in the policy, information from which to determine a millions of service unit (MSU) entitlement value and a cost entitlement value, the method comprising:

repeating on a periodic basis;

determining a pool of service units from billing entities that have respective rolling averages under respective MSU entitlement values;

determining a pool of billing entities that have respective rolling averages above respective MSU entitlement values, each billing entity in the pool having a priority assigned;

determining a cost pool based on unused cost entitlement units from billing entities that have respective rolling averages under respective MSU entitlement values;

determining, from the pool of billing entities, a favored billing entity that has a highest priority and a lowest deficit ratio;

adjusting the capacity limit for the favored billing entity above the MSU entitlement value with borrowed MSUs from the pool of service units, the borrowed MSUs being bounded by the cost pool;

scheduling work on the favored billing entity within the adjusted capacity limit, the scheduled work being executed on the favored billing entity; and

repeating determining a favored billing entity and adjusting until either the pool of service units is empty or the cost pool is empty.

9. The method of claim 8, wherein determining the pool of service units includes:

initializing the cost pool to a cost limit for the dynamic capping policy;

for billing entities having a respective rolling average greater than or equal to the respective MSUs entitlement, subtracting the product of the respective

MSU entitlement and an MSU cost from the cost pool; and

for billing entities having a respective rolling average less than the respective MSU entitlement, subtracting the product of the respective rolling average plus a buffer and a respective MSU cost from the cost pool.

10. The method of claim 9, wherein the MSU cost represents a relative value.

11. The method of claim 9, wherein the MSU cost represents a value calculated based on a prior bill for the mainframe.

12. The method of claim 8, wherein determining the pool of service units includes:

initializing the service unit pool to a MSU limit for the dynamic capping policy;

for billing entities having a respective rolling average greater than or equal to the respective MSU entitlement, subtracting the respective MSU entitlement from the service unit pool; and

for billing entities having a respective rolling average less than the respective MSU entitlement, subtracting the respective rolling average plus a buffer from the service unit pool.

13. A method comprising:

repeating on a periodic basis;

obtaining workload data from each of a plurality of billing entities identified in a dynamic capping policy, a billing entity being a logical partition of a mainframe computing system or a group of such logical partitions, each billing entity having a respective capacity limit, and the workload data for each billing entity including a respective rolling average of millions of service units (MSUs);

determining an entity pool, a service unit pool, and a cost pool for the plurality of billing entities based on, for each billing entity, the respective rolling average MSUs, a respective MSU entitlement value, and a respective MSU cost, the cost pool representing cost entitlement unused from billing entities that have respective rolling averages that are less than respective MSU entitlement values; and

dynamically adjusting capacity limits of the plurality of billing entities based on, the entity pool, the service unit pool, and the cost pool, wherein a workload manager schedules work within the respective capacity limits of the plurality of billing entities, the work being executed without exceeding a maximum cost limit and a maximum MSU limit, the maximum cost limit and maximum MSU limit applying to the plurality of billing entities.

14. The method of claim 13, wherein adjusting the respective capacity limits includes:

determining a favored billing entity of the plurality of billing entities that has a rolling average greater than its respective MSU entitlement value and a lowest low-importance workload percentage; and

adjusting the capacity limit for the favored billing entity above the MSU entitlement value using borrowed MSUs, the borrowed MSUs being bounded by the service unit pool and the cost pool.

15. The method of claim 14, wherein as part of determining the entity pool, the method includes:

determining that a first billing entity has a rolling average less than the MSU entitlement;

setting, responsive to the determining, the capacity limit for the first billing entity to the rolling average plus a buffer value;

adding the difference between the capacity limit and the MSUs entitlement to the service unit pool; and

adding the product of the MSU cost and the difference between the capacity limit and the MSU entitlement to the cost pool.

16. The method of claim 14, wherein bounding the borrowed MSUs includes:

determining a borrowed cost for the borrowed MSUs, the borrowed cost being equal to the product of the borrowed MSUs and an MSU cost for the favored billing entity;

determining whether the borrowed cost is greater than the cost pool;

reducing the cost pool by the borrowed cost when the borrowed cost is not greater than the cost pool; and

reducing the borrowed MSUs when the borrowed cost is greater than the cost pool.

17. The method of claim 13, wherein the MSU cost for a billing entity represents a relative value.

18. The method of claim 13, wherein the MSU cost for a billing entity represents a value calculated based on a prior bill for the mainframe computing system.

19. The method of claim 13, wherein adjusting the respective capacity limits includes:

determining that a billing entity that has a rolling average less than MSU entitlement; and

setting the respective capacity limit to the rolling average plus a buffer.