Abstract: Data or one or more operations can be provided, based on one or more characteristics associated with the data and/or operations, to a designated computing group or pool of computing resources designated for handling the data and/or operations with the particular data characteristic(s). The designated computing group can, for example, be one of multiple computing groups in the same system or device. As such, all of the computing groups can still function together in the same system or device, for example, in parallel. However, each one of the multiple computing groups can, for example, be defined or predefined to include one or more computing resources that are more suitable for storing and/or processing data with one or more data characteristics or handle operations with one or more determined characteristics.

Abstract: A system, method, and computer-readable medium that facilitate efficient use of cache memory in a massively parallel processing system are provided. A residency time of a data block to be stored in cache memory or a disk drive is estimated. A metric is calculated for the data block as a function of the residency time. The metric may further be calculated as a function of the data block size. One or more data blocks stored in cache memory are evaluated by comparing a respective metric of the one or more data blocks with the metric of the data block to be stored. A determination is then made to either store the data block on the disk drive or flush the one or more data blocks from the cache memory and store the data block in the cache memory. In this manner, the cache memory may be more efficiently utilized by storing smaller data blocks with lesser residency times by flushing larger data blocks with significant residency times from the cache memory.

Abstract: A system, method, and computer-readable medium that facilitate efficient use of cache memory in a massively parallel processing system are provided. A residency time of a data block to be stored in cache memory or a disk drive is estimated. A metric is calculated for the data block as a function of the residency time. The metric may further be calculated as a function of the data block size. One or more data blocks stored in cache memory are evaluated by comparing a respective metric of the one or more data blocks with the metric of the data block to be stored. A determination is then made to either store the data block on the disk drive or flush the one or more data blocks from the cache memory and store the data block in the cache memory. In this manner, the cache memory may be more efficiently utilized by storing smaller data blocks with lesser residency times by flushing larger data blocks with significant residency times from the cache memory.

Abstract: A system includes a multi-system database management system having a plurality of database systems. An index selection subsystem combines sets of query information from respective ones of the plurality of database systems into a workload. The index selection subsystem then generates candidate indexes from the workload, and selects recommended indexes from the candidate indexes based on one or more criteria.

Abstract: The number of writes to a device (“D1-n”) consisting of n sub-devices (“Da”) is counted: [ W 1 , 1 … W 1 , m - 1 W 1 , m W 2 , 1 … W 2 , m - 1 W 2 , m … … … … W n , 1 … W n , m - 1 W n , m ] ? ? where: Wa,b is the number of writes to sub-device Da in time period “b”; n>1; m>1. A write rate is determined for each sub-device Da at the end of time period p (“WRa,p”) using the following equation: WR a , p = ? b = 1 p ? ? W a , b ? b = 1 p ? ? t b ? where: tb is the amount of time in time period “b”; p<m. The write rate to sub-device Da is controlled in time periods after time period p so that device D1-n will have a guaranteed lifetime (“LD1-n”).

Abstract: Recovering from a failure of a soon-to-fail database system among a plurality of database systems in a multi-database system in processing a request submitted to the multi-database system through a multi-database system session is accomplished. A query band is created for the request. The query band is defined to be an identifier that uniquely identifies the session and the request among the plurality of sessions and plurality of requests being processed by the multi-database system at the time that the request was submitted. The query band is attached to a transaction issued by the multi-database system to the soon-to-fail database system to execute the query. The status concerning execution of the transaction is reported by the soon-to-fail database system. The status includes the query band. The status is logged. The failure of the soon-to-fail database system is detected.

Abstract: A system includes a multi-system database management system having a plurality of database systems. An index selection subsystem combines sets of query information from respective ones of the plurality of database systems into a workload. The index selection subsystem then generates candidate indexes from the workload, and selects recommended indexes from the candidate indexes based on one or more criteria.

Abstract: A technique for managing storage of a data object in a storage device involves receiving the data object (A) to store in the storage device, where the data object has an indicator bit pattern (P). Successive compression data transformations are applied to data object A to obtain respective corresponding compressed data objects, and one of these compressed data objects is selected, such that the selected compressed data object (C) has the shortest length with respect to the remaining compressed data objects. Compression information (I) is then associated with the compression data transformation used to generate data object C, and a threshold value T is calculated at least partly from the length of compression information I. If length (C)+T.gtoreq.length (A), then the indicator bit pattern of data object A is reset and the data object A is written to the storage device.

Abstract: A system, method, and computer readable medium are provided for reducing response time variation in a workload management system for a database system. When a query response is generated in response to a database query from a client, a determination may be made as to whether response time of the query response is less than a predetermined amount of time. Delivery of the query response to the client is delayed when an amount of time associated with the query response is less than the predetermined amount of time.

Abstract: A method and system for allocating storage space from storage devices within a computer system. The storage devices are grouped into respective homogeneous sets in which identical storage devices are included in the same homogeneous set. In accordance with the method and system, following receipt of an allocation request for storage space one or more stored response values are checked, the response values associated with at least one homogeneous set. The stored value or values represent the average response time of storage devices within the set. The homogenous set with the lowest average response time is selected. One or more stored allocation values are then checked that are associated with at least one storage device within the selected homogeneous set. The stored allocation value or values represent the current allocations of the associated storage device.

Abstract: Execution of a query is optimized. A plurality of plans for executing the query on a multi-database system are analyzed. A first plan having a first step and a second step is selected. The first step of the first plan is executed. The plan selection is re-evaluated before executing the second step of the first plan. A second plan of the plurality of plans for executing the query is executed to produce a result. The result is stored.

Abstract: Capacity of a computing environment that includes a database can be maintained at a target capacity by regulating the usage of one or more of the resources by one or more tasks or activities (e.g., database work). Moreover, the usage of the resource(s) can be regulated based on the extent of use of the resource(s) by one or more other activities not being regulated (e.g., non-database activities that cannot be regulated by a database system). In other words, a target capacity can be maintained by effectively adjusting the extent by which one or more tasks can access one more resources in consideration of the extent by which one or more of the resources are used by one or more other tasks or activities that are not being regulated with respect to their access of the resource(s).

Abstract: Computing capacity of a computing environment can be managed by controlling it associated processing capacity based on a target (or desired) capacity. In addition, fine-grained control over the processing capacity can be exercised. For example, a computing system can change the processing capacity (e.g., processing rate) of at least one processor operating based on a target capacity. The computing system may also be operable to change the processing capacity based on a measured processing capacity (e.g., a measured average of various processing rates of a processor taken over a period of time when a processor may have been operating at different processing rates over that period). By way of example, the processing rate of a processor can be switched between 1/8 and 2/8 of a maximum processing rate to achieve virtually any effective processing rates between them.

Abstract: A virtual regulator monitors and manages a plurality of database systems in a domain. Each of the database systems is tuned for a particular type of workload, and the virtual regulator or multiple virtual regulators running in parallel routes a set of one or more queries to a particular database system within the domain based on a cost function for each database system.

Abstract: A computer-implemented apparatus, method, and article of manufacture manage a plurality of database systems and perform data maintenance tasks in a data warehouse system. A domain includes a plurality of database systems. A virtual regulator manages the domain, detects a request to invoke a data maintenance task on a first system in the domain, routes the data maintenance task, for execution, to a second system in the domain, and applies results from the data maintenance task (executed by the second system) to the first system.

Abstract: Techniques for selective compression of database information are presented. Selective first portions of a field associated with a database table are identified along with metadata associated with the database table. The first portions and the metadata are then compressed by selecting one or more compression algorithms in response to compression capabilities associated with the algorithms.

Abstract: A system for managing a plurality of systems, includes a domain comprised of a plurality of systems, wherein the systems are managed by one or more closed-loop system management (CLSM) regulators. A virtual regulator manages operational characteristics of the domain and the plurality of systems, by communicating with the CLSM regulators to monitor the systems' states, events and conditions within the domain and to adjust the systems' operational characteristics in response to the systems' states, events and conditions.

Abstract: Query processing statistics are logged in a multi-database system containing a plurality of system databases. A plurality of query log entries generated using information regarding the execution of a plurality of queries are temporarily stored in a database query log (DBQL). The information is received from a plurality of system databases. Each query log entry is generated using information regarding execution of one query by the plurality of system databases. Each query log entry includes a field identifying a system database that was the source of the information in that entry. An express request is created upon occurrence of a triggering event. The express request contains a subset of the query log entries temporarily stored in the DBQL cache. The created express request is transmitted to a plurality of system databases and the contents of the transmitted express request are stored in each of plurality of system databases to which it was transmitted.

Abstract: Optimizing the execution of a query in a multi-database system includes identifying a region within a table, the table being referenced in the query. The region is stored on data-storage devices on first and second system databases in the multi-database system. A first access plan for the query is developed, the first access plan comprising accessing the version of the region stored on the first system database. A second access plan for the query is developed, the second access plan comprising accessing the version of the region stored on the second system database. A selection is made between the first access plan and the second access plan to execute the query. The query is executed using the selected access plan to produce a result.

Abstract: Optimizing the execution of a query in a multi-database system includes identifying a region within a table, the table being referenced in the query. The region is stored on a data-storage device on a first of the system databases in the multi-database system. The region is stored on a data-storage device on a second of the system databases in the multi-database system, the second system database being a different system database than the first system database. A first access plan for the query is developed, the first access plan comprising accessing the version of the region stored on the first system database. A second access plan for the query is developed, the second access plan comprising accessing the version of the region stored on the second system database. A selection is made between the first access plan and the second access plan to execute the query. The query is executed using the selected access plan to produce a result. The result is stored.