Abstract: Efficient buffer cache utilization frees a data buffer as soon as data buffer processing is completed, and without losing association of the freed data buffer and a descriptor buffer. Separate free buffer link lists identify the freed data buffer and any freed descriptor buffer. The data buffer is rapidly processed then freed generally before completion of processing of the descriptor buffer, freeing the processed associated data buffer before the associated descriptor buffer is freed. The association of the processed free data buffer and the descriptor buffer may be ended to enable the more frequent use of the large capacity data buffer for other update requests.

Abstract: A method to access an object in a computer system storage includes receiving a reference to the object, obtaining a referential map based on the reference (the referential map indicating a plurality of logical storage locations), each logical storage location associated with a portion of the object, and obtaining a physical map that is independent of the referential map from one of the logical storage locations indicated in the referential map, the physical map indicating a physical storage location of at least a part of that portion of the object being accessed. Computer storage systems providing distributed storage in accordance with this method and methods storage media having instructions to perform the method are also described.

Abstract: Efficient buffer cache utilization frees a data buffer as soon as data buffer processing is completed, and without losing association of the freed data buffer and a descriptor buffer. Separate free buffer link lists identify the freed data buffer and any freed descriptor buffer. The data buffer is rapidly processed then freed generally before completion of processing of the descriptor buffer, freeing the processed associated data buffer before the associated descriptor buffer is freed. The association of the processed free data buffer and the descriptor buffer may be ended to enable the more frequent use of the large capacity data buffer for other update requests.

Abstract: An electronic storage system, such as a file system, may include a storage access routine to store data objects. Data objects stored in the system have an entry in an object index that may be maintained by the storage access routine. The object index includes fault tolerance metadata for each data object that specifies the fault tolerance techniques used in storing a data object, such as a particular RAID level. The fault tolerance metadata maintained for each object allows different data objects to be stored with different fault tolerance techniques on the same storage volume. Additionally, the metadata for a data object may be modified in order to store the object using different fault tolerance techniques.

Abstract: A controller fault recovery system recovers from faults that cause unscheduled stops for a distributed file system operating on an array storage system having multiple controllers. A proxy arrangement protects data integrity in the event of an unscheduled stop on just one controller in the array storage system. An atomic data/parity update arrangement protects data integrity in the event of an unscheduled stop of more than one controller.

Abstract: A first file system configuration may be dynamically migrated (morphed) into a second file system configuration on an object-by-object basis. During the migration operation, objects (directories and files, data and metadata) may be accessed by client applications through, either the first or second file system configurations. That is, objects are transparently shared between co-active first and second file system configurations. Additionally, a file system morph operation already in progress may be canceled with all objects previously morphed to the second file system configuration returned to the first file system configuration.

Abstract: A self-healing computer storage system utilizes a proxy storage management process to service memory access requests directed to stored objects whose designated storage management process has failed. The proxy accesses the relevant parts of the stored objects fault tolerance information to service memory access requests, updating the stored object's fault tolerance information to reflect any changes. When the previously failed storage management process is restarted, it determines if the fault tolerance information for any of the objects (or parts thereof) it manages have been modified (i.e., by a proxy). If such indication is found, the restarting storage management process reconstructs its stored object data (and metadata) from the stored objects' fault tolerance information.

Abstract: A method to manage storage of an object in a computer system having a first and a second storage management process (wherein the stored object has a data portion, a metadata portion and a fault tolerance data portion) includes receiving a memory access request from a client process, routing the memory access request to the first storage management process, determining the first storage management process has failed, routing the memory access request to the second storage management process (having access to the fault tolerance data portion), receiving a result from the second storage management process, and returning at least a portion of the result to the client process. The second storage management process may reconstruct at least a portion of the metadata portion, modify the fault tolerance data portion in accordance with the memory access request, and store the modified fault tolerance information.

Abstract: The invention discloses apparatus and process in which data files are distributed across a large scale data processing system to enable protection from the loss of data due to the failure of one or more fault domains. Specifically, the invention provides significant advances in data base management by distributing data across N fault domains using one or more of a multitude of deterministic functions to protect failure.

Abstract: The invention discloses apparatus and process in which data files are distributed across a large scale data processing system to enable balance of work loads and storage loads at a plurality of nodes. Specifically, the invention provides significant advances in data base management by distributing meta-data in a plurality of file storage nodes to isolatively and distributively store file data in a distributed computing environment. This provides efficient allocation of storage space and work loads among nodes.

Abstract: A distributed network is described which contains a plurality of interconnected nodes each node including a processor and data storage apparatus. A plurality of key-index data identifiers are distributed among the nodes, with each node including a tree data structure in non-volatile storage defining locations of the key-index data identifiers. The tree data structure includes a ROOT data structure comprising two lists, "NEW ROOT" and "OLD ROOT", each comprised of an ordered array of boundaries assigned nodes for the top level of the tree. A method is described for balancing the tree data structure which comprises the steps of:a. a providing in each of the nodes across which the key-index data identifiers are distributed, at least copies of the two lists, "NEW ROOT" and "OLD ROOT", of the ROOT data structure;b. determining when a first node contains an excess of key-index data identifiers;c. moving the excess of key-index data identifiers to a second node;d.

Abstract: A data processing system includes both volatile and nonvolatile memories, user-inaccessible file control software and performs a method for recovering from a malfunction occurring during a file transaction that prevents accomplished microscopic transactions from being lost.

Abstract: A multi-dimensional, multi-nodal routing mechanism is described for relaying information from node to node using a header consisting of route descriptor bits. Each node's receiver/transmitter pair changes states as the information is guided to the destination node. The message is propagated over several nodes simultaneously to traverse the nodes and reach the destination node quickly. When the final node is reached, all alternate communication routes are freed.

Abstract: This is method for parity protecting distributed data files in a multi-node, parallel data processing network, with each node having a data store. Each data file is organized as a series of fixed size file data blocks, each file data block being comprised of a block identifier and a plurality of data bit positions. The method comprises: distributing at least N-1 file data blocks across N-1 nodes, each of the N-1 file data blocks being stored in a data store at an available physical location, which physical location may be selected from any within an area of the data store reserved for file data blocks, each data store having an associated directory indicating where each file data block is located; exclusive OR'ing data bits in corresponding bit positions in each of the N-1 file data blocks; and storing the Exclusive OR results in corresponding bit positions of a parity file data block in a data store in node N.

Abstract: The method of the invention is performed on a disk drive with at least a rotatable disk and a head/arm mechanism movable thereover. The disk has a plurality of concentric tracks, including innermost and outermost tracks and an intermediate track positioned therebetween, with the intermediate track including often-used records. The method reduces movements of the head/arm mechanism through the steps of: positioning the head/arm mechanism over the intermediate track; moving the head/arm mechanism away from the intermediate track to another track to perform a requested read operation; moving the head/arm mechanism back to the intermediate track from the further track and performing write actions in the tracks over which the head/arm mechanism moves in its traversal back to the intermediate track.

Abstract: An improved multiplier is disclosed for multiplying a first operand times a second operand, which includes a Booth-type translator having an input connected to receive the first operand, for translating the binary expression of the first operand into a sequence of signed digits. The multiplier further includes a partial product generator having a first input connected to the output of the translator and a second input connected to receive the second operand, for multiplying the translated first operand times the second operand and outputting partial products consisting of signed digits. The multiplier further includes an array of adders, each adder having an input connected to two of the signed digits output from the partial product generator, for providing a sum consisting of a sequence of signed digits.

Abstract: An instruction pipeline for a data processor is disclosed, in which instruction execution is carried out in a sequence of phases which include fetching the instruction from an instruction storage, computing a data storage address from the fetched instruction, accessing the data storage at the computed address to obtain a datum operand, and then carrying out the logical or arithmetic operation on the accessed datum in accordance with the fetched instruction. Branch and stack instructions and return instructions are accommodated by providing a return address stack in the data storage, which stores the next instruction store address to be returned to after a return operation has been completed.