International Business Machines Corporation

Erik Alan Daughtrey

Gennaro A Cuomo

Brian Keith Martin

Matt Richard Hogstrom

Christopher J Blythe

Yen-Fu Chen

Asser Nasreldin Tantawi

Malgorzata Steinder

Nikolaos Anerousis

Giovanni Pacifici

Michael Joseph Spreitzer

Michael L Fraenkel

Hari Shankar

Wenjian Qiao

John S Cox

Incoming work units (e.g., requests) in a computing workload are analyzed and classified according to predicted execution. Preferred embodiments track which instrumented wait points are encountered by the executing work units, and this information is analyzed to dynamically and autonomically create one or more recognizers to programmatically recognize similar, subsequently-received work units. When a work unit is recognized, its execution behavior is then predicted. Execution resources are then allocated to the work units in view of these predictions. The recognizers may be autonomically evaluated or tuned, thereby adjusting to changing workload characteristics. The disclosed techniques may be used advantageously in application servers, message-processing software, and so forth.

Autonomic workload classification using predictive assertion for wait queue and thread pool selection

The present invention provides a method, system, and computer program product for supporting a large number of intermittently used application clusters without consuming all the resources needed to run server processes for all the application clusters all the time. A method in accordance with the present invention comprises: holding an application request for a dormant application cluster in a holding zone; changing a status of the dormant application cluster to active; changing a status of an unused active application cluster to dormant; stopping server processes for the newly dormant application cluster; starting server processes for the newly active application cluster; and directing the application request held in the holding zone to the server processes for the newly active application cluster.

Method, system, and computer program product for supporting a large number of intermittently used application clusters

Methods, systems, computer program products, and methods of doing business whereby programmatically-generated byte code insertion is used to perform run-time tracing of code that potentially encounters a wait during execution. The byte code insertion is performed at load time, and inserts byte codes before and after a located (potential) wait point. The inserted byte code functions to gather execution statistics, such as a time stamp before invoking a located wait point and a time stamp after invoking the located wait point. Preferred embodiments allow this tracing to be selectively activated/deactivated.

Run-time wait tracing using byte code insertion

The invention described is a new and useful process for minimizing the overall rebate a provider disburses to customers when a service level agreement (SLA) breach occurs in a utility computing environment. Specifically, the process compares performance data and resource usage with the SLAs of the customers, and reallocates shared resources to those customers who represent a lesser penalty to the provider in the event of an SLA breach. The process determines which resources, used by customers representing the lesser penalty, are operating below peak capacity. The process then reallocates these under-utilized resources to those customers requiring additional resources to meet SLA thresholds. If all resources are operating at peak capacity, the process reallocates the resources to those customers whose SLAs represent a greater penalty in the event of an SLA breach as compared to those customers whose SLAs provide for a lesser penalty, thereby minimizing the total rebate due upon an SLA breach.

Algorithm for minimizing rebate value due to SLA breach in a utility computing environment

Thread pools in a multithreaded server are programmatically adjusted, based on observed statistics from the server's inbound workload. In a multithreaded server environment, response time to end users is improved while increasing the efficiency of software execution and resource usage. Execution time and wait/queued time are tracked, for various types of requests being serviced by a server. Multiple logical pools of threads are used to service these requests, and inbound requests are directed to a selected one of these pools such that requests of similar execution-time requirements are serviced by the threads in that pool. The number and size of thread pools may be adjusted programmatically, and the distribution calculation (i.e., determining which inbound requests should be assigned to which pools) is a programmatic determination. In preferred embodiments, only one of these variables is adjusted at a time, and the results are monitored to determine whether the effect was positive or negative. The disclosed techniques also apply to tracking and classifying requests by method name (and, optionally, parameters).

Dynamic thread pool tuning techniques

Workload is programmatically distributed across a set of execution resources. In a multithreaded server environment, response time to end users is improved while increasing the efficiency of software execution and resource usage. Execution time and wait/queued time are tracked, for various types of requests being serviced by a server. Multiple logical pools of threads are used to service these requests, and inbound requests are directed to a selected one of these pools such that requests of similar execution-time requirements are serviced by the threads in that pool. The number and size of thread pools may be adjusted programmatically, and the distribution calculation (i.e., determining which inbound requests should be assigned to which pools) is a programmatic determination. In preferred embodiments, only one of these variables is adjusted at a time, and the results are monitored to determine whether the effect was positive or negative. The disclosed techniques also apply to tracking and classifying requests by method name (and, optionally, parameters).

Programmatic response-time based workload distribution techniques

An object state caching method can include augmenting a compiled object having at least one referenced method with conditional caching logic. The conditional caching logic can be executed in an execution environment to determine whether to execute the referenced method. If the conditional caching logic determines not to execute the referenced method, the execution of the referenced method can be bypassed. Additionally, the stored object state can be retrieved from a communicatively coupled object cache. Finally, the retrieved object state can be forwarded to a requesting process. If, however, the conditional caching logic determines to execute the referenced method, the referenced method can be executed producing the resultant object state, and, the resultant object state can be cached in the coupled object cache.

Growing community of inventors

Durham, NC, United States of America

Erik Alan Daughtrey