Method for supporting on-demand performance

• US 7,729,270 B2
• Filed: 01/13/2005
• Issued: 06/01/2010
• Est. Priority Date: 01/13/2005
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Johnson, Randy Scott, Kenley, James Michael, Rinckel, H. William, Bishop, Ellis Edward, Northway, Tedrick Neal, Zolotow, Clea Anne
• Primary Examiner(s): Ngo; Ricky
• Assistant Examiner(s): OVEISSI, MANSOUR
• Application Number:
  US11/034,905
• Publication Number:
  US 20060153090A1
• Time in Patent Office: 1,965 Days
• Field of Search: 370/252, 370/241, 370/242, 370/229, 709/203, 709/224
• US Class Current: 370/252
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04L 67/61 taking into account QoS or priority requirements

Abstract

A method is disclosed for managing the performance of shared computing resources in furtherance of delivering on-demand IT services to multiple customers. The inventive method comprises predicting the performance of resources allocated among the customers to minimize performance related problems; preventing performance related problems by evaluating the impact of changes in resource allocations; reacting to performance alerts to resolve performance related problems; and determining the appropriate level of resource performance needed in service level agreements, and that these requirements detailed in the service level agreements are achieved. This Performance Management method ensures that the demands of the customers are not affected by the allocation and utilization of the shared IT resources.

Claims

1. A method for managing the performance of an on-demand, shared computing system, at the system and subsystem level, the method comprising:

managing the performance of the on-demand, shared computing system by a computer having at least one processor executing programming instructions to perform steps comprising;

monitoring a plurality of performance data;

analyzing the plurality of performance data to identify a performance trend;

identifying an approaching performance degradation in the performance trend;

identifying a component of the on-demand, shared computing system that the approaching performance degradation impacts;

identifying a performance parameter of the on-demand, shared computing system that impacts the component;

determining if a system limitation constrains an adjustment of the performance parameter;

responsive to determining that the system limitation does not constrain the adjustment of the performance parameter, comparing the performance trend with a hypothetical optimum performance trend to identify a correct adjustment to the performance parameter needed to return the component to acceptable performance;

creating a recommendation indicating the correct adjustment to the performance parameter, and forwarding the recommendation to an appropriate party;

responsive to receiving approval of the recommendation from the appropriate party, implementing the recommendation to resolve the approaching performance degradation; and

responsive to determining that the system limitation does constrain the adjustment of the performance parameter, communicating the system limitation to the appropriate party, and opening a change request to resolve the limitation.

2. A method for reacting to an alert in a shared computing system, the process comprising:

reacting to the alert in the shared computing system by a computer having at least one processor executing programming instructions to perform steps comprising;

determining if the alert is a performance alert;

responsive to determining the alert is not a performance alert identifying a service group that should receive the alert, and routing the alert to the service group;

responsive to determining that the alert is a performance alert, determining if the performance alert is valid;

responsive to determining the performance alert is invalid, creating a documentation of the performance alert being invalid and forwarding the documentation to a Performance Analyst;

responsive to determining the performance alert is valid, determining if additional data is required to evaluate the performance alert;

responsive to determining that additional data is required to evaluate the performance alert, generating an ad hoc report containing the additional data;

identifying a component of the shared computing system that caused the performance alert,

identifying a variable in the component that caused the performance alert,

modifying the variable to resolve the performance alert, and

monitoring the component to ensure the performance alert is resolved.

3. A process for managing the performance of resources in a shared computing system to meet requirements of a contract governing a service level agreement, the process comprising:

reviewing the contract, the service level agreement, and a technology strategy, wherein the service level agreement is stored in an object-oriented database;

identifying a performance parameter to be monitored for ensuring that the service level agreement is satisfied;

establishing a performance baseline for the performance parameter;

defining a performance threshold, based on the performance parameter and the performance baseline, for alerting a Performance Analyst when a performance parameter crosses the performance threshold;

gaining approval for the performance threshold from an account team to ensure the service level agreement satisfies the requirements of the contract;

defining a performance alert and a response to the performance alert;

incorporating the defined performance alert, the performance threshold, and the response into the service level agreement; and

monitoring the system performance to ensure the requirements of the service level agreement are satisfied, by a computer having at least one processor executing programming instructions to perform steps comprising;

establishing data collection requirements;

collecting and reporting data according to the data collection requirements;

receiving and reviewing the data collected; and

changing the performance alert and the performance threshold to meet the requirements of the service level agreement.