Abstract: A storage system includes a plurality of data vats, and a processor including an optimizing unit that optimizes a value of data stored in the storage system. The optimizing unit optimizes the value by computing and implementing an optimal decision for allocating new data to a first data vat of the plurality of data vats, moving existing data from at least a second data vat of the plurality of data vats to the first data vat, and deleting existing data from the first data vat, based on an amount of data in each of the plurality of data vats.

Abstract: Provided are a method, system, and program for managing locks enabling access to a shared resource. A first server receives a lock request from a client for the shared resource. A determination is made as to whether a second server owns the client locks. The first server issues a request to the second server to transfer ownership of the client locks to the first server, wherein the client lock requests are handled by the server owning the client locks.

Abstract: A storage system includes a plurality of data vats, and a processor including an optimizing unit that optimizes a value of data stored in the storage system. The optimizing unit optimizes the value by computing and implementing an optimal decision for allocating new data to a first data vat of the plurality of data vats, moving existing data from at least a second data vat of the plurality of data vats to the first data vat, and deleting existing data from the first data vat, based on an amount of data in each of the plurality of data vats.

Abstract: A method (and system) of storing data in a value-based storage system, includes optimizing a value of data stored in the value-based storage system.

Abstract: A method for reducing data loss and unavailability by integrating multiple levels of a storage hierarchy is provided. The method includes receiving a read request. In addition, the method includes recognizing a data failure in response to the read request. The method further includes locating an alternate source of the data to be read in response to recognizing the data failure. The alternate source includes data cached at devices in the storage hierarchy, data in a backup system, and cumulative changes to the data since the last backup. Moreover, the method includes responding to the read request with data from the alternate source.

Abstract: A system and method for optimizing a storage system to support short data object lifetimes and highly utilized storage space are provided. With the system and method, data objects are clustered based on when they are anticipated to be deleted. When an application stores data, the application provides an indicator of the expected lifetime of the data, which may be a retention value, a relative priority of the data object, or the like. Data objects having similar expected lifetimes are clustered together in common data structures so that clusters of objects may be deleted efficiently in a single operation. Expected lifetimes may be changed by applications automatically. The system automatically determines how to handle these changes in expected lifetime using one or more of copying the data object, reclassifying the container in which the data object is held, and ignoring the change in expected lifetime for a time to investigate further changes in expected lifetime of other data objects.

Abstract: A method for reducing data loss and unavailability by integrating multiple levels of a storage hierarchy is provided. The method includes receiving a read request. In addition, the method includes recognizing a data failure in response to the read request. The method further includes locating an alternate source of the data to be read in response to recognizing the data failure. The alternate source includes data cached at devices in the storage hierarchy, data in a backup system, and cumulative changes to the data since the last backup. Moreover, the method includes responding to the read request with data from the alternate source.

Abstract: When an alarm condition relating to a performance goal of a storage system is detected, a storage management system invokes an N-step lookahead engine for simulating operation of the storage system when there are multiple actions that could be taken by the storage system for eliminating the alarm condition. The N-step lookahead engine generates N possible system states based on a current state of the storage system. The N possible states are based on a cost model of each of the multiple actions. Each cost model is based on an action, a behavior implication of the action, a resource implication of the action and a transient cost of the action. An action is selected that generates a system state that optimizes the stability, a prerequisite and a transient cost of invoking the selected action.

Abstract: A method for reducing data loss and unavailability by integrating multiple levels of a storage hierarchy is provided. The method includes receiving a read request. In addition, the method includes recognizing a data failure in response to the read request. The method further includes locating an alternate source of the data to be read in response to recognizing the data failure. The alternate source includes data cached at devices in the storage hierarchy, data in a backup system, and cumulative changes to the data since the last backup. Moreover, the method includes responding to the read request with data from the alternate source.

Abstract: A method (and system) of storing data in a value-based storage system, includes optimizing a value of data stored in the value-based storage system.

Abstract: In a multiprocessing system, access to a shared resource is arbitrated among multiple computing nodes. The shared resources has a membership view resulting from a predetermined membership protocol performed by the shared resource and the computing nodes. Preferably, this membership protocol includes a termination condition guaranteeing asymmetric safety among all members of the multiprocessing system. The shared resource arbitrates access to itself by fencing computing nodes outside shared resource's membership view. In one embodiment, the shared resource may comprise a data storage facility, such as a disk drive. Illustratively, computation of the shared resource's membership view may employ a procedure where each computing node subscribes to the resource during prescribed membership intervals.

Abstract: Provided are a method, system, and program for managing locks enabling access to a shared resource. A first server receives a lock request from a client for the shared resource. A determination is made as to whether a second server owns the client locks. The first server issues a request to the second server to transfer ownership of the client locks to the first server, wherein the client lock requests are handled by the server owning the client locks.

Abstract: A method and apparatus for compression and decompression of information, such as groups of computer program instructions, encodes (compresses) information comprising a plurality of units by receiving the information to be encoded, splitting the information into a plurality of subsets, each subset comprising a plurality of symbols, each symbol comprising at least a portion of a unit of information, and assigning a codeword to each symbol, for each subset. Preferably, the assignment is performed by determining the frequency of occurrence of each symbol, for each subset, and assigning a codeword to each symbol, based on the frequency of occurrence of each symbol, for each subset. In order to decode (decompress) encoded information, the information comprising a plurality of codewords, each codeword is decoded to form a symbol, each symbol is grouped into one of a plurality of subsets and the plurality of subsets is merged to form decoded information.

Abstract: A disk controller system includes a microprocessor, a hard disk controller, a disk channel path, a host communications path, and an interface coupled to each of the microprocessor, hard disk controller, disk channel path and host communications path. A unified non-volatile memory is coupled to the interface that has a plurality of memory spaces. A memory space is allocated for each of the microprocessor, hard disk controller, disk channel path and host communications path. Each memory space is separated from another memory space by a programmable memory space boundary. The microprocessor, hard disk controller and the unified memory are all fabricated on a single substrate.

Abstract: In a multiprocessing system, access to a shared resource is arbitrated among multiple computing nodes. The shared resources has a membership view resulting from a predetermined membership protocol performed by the shared resource and the computing nodes. Preferably, this membership protocol includes a termination condition guaranteeing asymmetric safety among all members of the multiprocessing system. The shared resource arbitrates access to itself by fencing computing nodes outside shared resource's membership view. In one embodiment, the shared resource may comprise a data storage facility, such as a disk drive. Illustratively, computation of the shared resource's membership view may employ a procedure where each computing node subscribes to the resource during prescribed membership intervals.

Abstract: A method, apparatus, article of manufacture, and a memory structure for storing and retrieving data in physical sectors which are larger than the sector size presented to the user is disclosed. The method comprises the steps of receiving at least one user sector comprising write data, and writing the user sector to a portion of the physical sector of the storage device. In one embodiment, data is written to the data storage disk using existing data that was read from the storage disk for a RAID parity calculation.

Abstract: An apparatus and method for transparent on-the-fly decompression of the program instruction stream of a processor. Connected between a processor and a memory storing compressed information is a decompression device. The decompression device, receives a request from the processor for information, retrieves compressed information from the memory, decompresses the retrieved compressed information to form uncompressed information, and transmits the uncompressed information to the processor. The compressed information may include both program instructions and data. When the decompression device receives a request for information, which includes an unmodified address, from the processor, it generates an index offset from the received unmodified address. An indexed address corresponding to the generated index offset is retrieved from an index table. Compressed information corresponding to the selected indexed address is retrieved from the memory and transmitted to the processor.

Abstract: Ordered machine-readable messages are reliably delivered among processing members in a multiprocessing computer system. The system includes multiple processing nodes, each having a unique source-ID and a membership view including one or more of the processing nodes with which it can nominally exchange messages. When a stimulus message is received by a first processing node, the node increments a coordinated local counter (CC). The node also sends a multicast message to all processing nodes in the first node's membership group. The multicast message includes the received stimulus message, the incremented CC value, and the first node's source-ID. The node further sets a timer, exclusively associated with the incremented CC value. When a multicast message is received at a processing node, the node performs a multicast input processing routine. The node sets its CC equal to the greater of its current value or the received multicast message's CC value.

Abstract: A method is disclosed for coordinating membership subject to an asymmetric safety condition from multiple processes in a distributed system. Each is callable by a distributed application for system status or for executing tasks of the application. Initially, each process sends the other processes its view on their status, where the view includes the names of a group of the processes. It then waits for similar views from other processors except those regarded as failed in its own view, up to a predetermined timeout. Each process then generates a resulting view by intersecting its local view with the names of those processes from which it has received views. The local views are updated based on the resulting views and again exchanged until a termination condition occurs.