Sas Institute Inc.

Matthew Scott Maxwell

Tugrul Sanli

Xiaodong Yao

Feng Chen, Ii

Feng (Jason) Chen

Jing (Jennie) Hu

Exemplary embodiments relate to the problem of allocating a finite number of units of a resource among requestors willing to offer different amounts of value for the resource. When different classes of requestors are permitted to cancel the request or fail to show up to collect the unit of the resource with different probabilities (collectively referred to as 'wash'), the problem becomes difficult to solve efficiently. According to the procedures described herein, the capacity is artificially inflated to offset the impact of wash, and then protection levels are computed using the inflated capacity as if there was no wash. The capacity is then artificially inflated again based on the new protection levels, and the process is repeated until, e.g., the results converge. Using this procedure, overallocation limits and protection levels can be computed in real-time, and accordingly the resource can be allocated efficiently as new requests are received.

Two-part job scheduling with capacity constraints and preferences

In one example, a system can receive configuration data indicating how resources can be combined, identify availability data indicating the total number of various resources that are available, and determine maximum-capacity data using the availability data and the configuration data. The system can also receive distribution data having probability distributions for jobs to be implemented using the resources, determine capacity valuations using the distribution data and the availability data, and determine a configuration of resources using the capacity valuations and the maximum-capacity data. Thereafter, the system can receive a job and determine that a valuation for the job exceeds a predefined threshold associated with using the configuration of resources. In response to determining that the valuation exceeds the predefined threshold, the system can assign the resources to the job in the configuration. The system can then cause the job to be performed using the configuration of resources assigned to the job.

Managing resources in a configurable computing environment

Resources can be managed by predicting resource usage. For example, a first computing resource having a first attribute with a first value can be identified. A second computing resource having a second attribute with a second value can also be identified. A configuration parameter can be determined based on at least one of historical data or a predetermined probability. A predetermined model can be used to predict a first amount of interest in the first computing resource based on the first value, the second value, and the configuration parameter. The predetermined model can also be used to predict a second amount of interest for the second computing resource based on the first value, the second value, and the configuration parameter. Information related to the predicted first amount of interest in the first computing resource and the predicted second amount of interest in the second computing resource can be transmitted.

Growing community of inventors

Cary, NC, United States of America

Matthew Scott Maxwell