Abstract: Embodiments of the invention provide a parcel-based, data-mapping scheme that allow for implementation of data integrity methods and variable size logical data blocks while the layout of the physical storage device remains unchanged. For one embodiment, the invention provides a method in which a virtual data storage parcel including a number of extended-size logical data storage blocks is created, and one or more physical data storage parcels, each including a number of standard-size logical data storage blocks, is created. The combined size of the one or more physical data storage parcels equals or exceeds the size of the virtual data storage parcel. The extended-size logical data storage blocks of the virtual data storage parcel are mapped to the standard-size logical data storage blocks of the one or more physical data storage parcels.

Abstract: Various embodiments of systems and methods for implementing a streaming I/O protocol are disclosed. In some embodiments, a method may involve: receiving a packet initiating a streaming write operation, where the packet indicates that the size of the streaming write is larger than the size of the packet; initiating a write access having a size larger than the size of the packet to a storage device; receiving subsequent packets included in the streaming write operation; and writing data received in the subsequent packets to the storage device as part of the write access initiated in response to the earlier packet. In some embodiments, streaming read operations may also be supported.

Abstract: Embodiments of the present invention provide methods and apparatuses for accessing MD atomically. For one embodiment, the local cache and the remote cache are each provided with an MBE. To perform an atomic access operation, the local processing system allocates a burst descriptor buffer. The burst descriptor buffer is used to specify the addresses to be atomically accessed. For write operations, the burst descriptor buffer also specifies the update data. The addresses, and data, if any, are then transmitted to the remote cache as a series of packets constituting a single data access request. The MBE of the remote cache then performs the operations contained in the request. Upon completion of all of the operations the remote MBE transmits an acknowledgement. For read operations the remote MBE also transmits the requested data.

Abstract: An embodiment of the invention provides a method for creating a snapshot of a data store. A command to create a snapshot of an original data store, stored to a parent virtual logical unit (VLU), is received. A deferred propagation data structure (DPDS) is associated with the parent VLU. The DPDS is capable of containing data propagation records and separators, each data propagation record contains a previous version of one or more data blocks of the parent VLU, each separator contains a pointer to a particular child VLU storing a snapshot of the original data store and separating the data propagation records pertaining to the particular child VLU. A child VLU to store the copy of the original data store is created. A new separator containing a pointer to the child VLU is created in the DPDS. A search pointer pointing to the DPDS is implemented in the child VLU.

Abstract: A system and related method are described for booting from one of multiple boot images that includes a processor coupled to a main memory, a number of boot devices and a selector. The boot devices are used to store multiple copies of a boot image and are in communication with the processor through a controller. The selector is used to select a different boot device each time the system is turned on or reset. A set of instructions executed by the processor loads the boot image from one of the boot devices selected by the selector into the main memory. In one embodiment, the selector is embodied in the form of software instructions incorporated into or linked to a system BIOS program. In another embodiment, the selector is embodied in the form of software instructions incorporated into or linked to a SCSI BIOS program.

Abstract: An embodiment of the invention provides a method for improving efficiency in a system containing repeated snapshots of an original data store. A READ/WRITE enabled parent virtual logical unit (VLU) to store an original data store is created. One or more READ-only enabled child VLUs, each storing a point-in-time copy of the original data store, are created at distinct times. A search pointer and a propagation pointer are implemented on the parent VLU and on each child VLU. The search pointer of each child VLU is set to point to the next subsequently created child VLU, with the search pointer of a most recently created child VLU set to point to the parent VLU. The propagation pointer of each child VLU is set to point to the immediately previously created child VLU, with the propagation pointer of the parent VLU set to point to the most recently created child VLU.

Abstract: A fault tolerant data storage system and related method are described for effectively failing over a storage device from one storage controller to another storage controller. The storage system generally includes at least two storage controllers for coupling to a processor and at least one storage device. A failover manager is in communication with the storage controllers and the storage device. The failover manager assists failing over of the storage device by allowing only one of the storage controllers having ownership to access the storage device at any one time. In one embodiment, the failover manager maintains a list of recent requests that have been committed to the storage device so that it can be used during failover to assist the surviving controller to complete the uncommitted requests properly.

Abstract: A method for creating a point-in-time copy of a data store. A command to create a point-in-time copy of an original data store is received. The processing of input/output operations to the original data store is suspended. A metadata structure of a copy data store is created. The metadata structure contains a null metadata slab tree, a backing store register, and a copy record, the copy record containing a pointer to a metadata structure of the original data store, the metadata structure of the original data store containing a pointer to the metadata structure of the copy data store. The processing of input/output operations to the original data store and to the copy data store is resumed.

Abstract: In one embodiment, a direct memory access (DMA) disk controller used in hardware-assisted data transfer operations includes command receiving logic to receive a data transfer command issued by a processor. The data transfer command identifies one or more locations in memory and multiple distinct regions on one or more disks accessible to the DMA disk controller. The DMA disk controller further includes data manipulation logic to transfer data between the memory locations and the distinct regions on the disks according to the data transfer command.

Abstract: A metadata tree structure having a plurality of nodes (slabs), each node containing a MD table. Each of the MD tables has a plurality of entries. Each of the entries in the MD tables represents a contiguous range of block addresses and contains a pointer to a cache slot storing a date block corresponding to the block address, or an indicator to indicate that the corresponding data block is not stored in a NVRAM cache slot. Each MD table also contains a block address range indicator to indicate the contiguous range of block addresses, and at least one pointer to point to any parent or child nodes. In an alternative embodiment, the pointer of each MD entry may point to a disk address if the data is not in cache. For such an embodiment, portions of the MD store may be stored to disk.

Abstract: A data storage system is disclosed in which a 3-party hand-off protocol is utilized to maintain a single coherent logical image. In particular, the functionality of the data storage system is separated into distinct processing modules. There are at least three types of processing modules. These three kinds of processing modules function cooperatively to perform data storage operations via the 3-party hand-off protocol. At least two types of processing modules are required to cooperate in a hand-off manner to fully execute each data storage operation.

Abstract: A data storage system is disclosed in which a 3-party hand-off protocol is utilized to maintain a single coherent logical image. In particular, the functionality of the data storage system is separated into distinct processing modules. Each processing module is implemented in a distinct central processing unit (CPU). Alternatively, the first type processing module and the third type processing module can be implemented in a common CPU. Isolating the different functions of the data storage system into distinct CPUs facilitates failure recovery. A characteristic of the 3-party hand-off protocol is that, if an abnormal state occurs, a surviving module has sufficient information to proceed to recover from said abnormal state after detecting the abnormal state, without depending on a failing module, by retrying the data storage operation with another processing module or the failing module or cleaning up after the failed data storage operation, resulting in improved failure recovery.

Abstract: A memory system is disclosed in which hardware assisted leases are utilized to control access to shared data resources among data users. The lease is a type of lock on content in the memory system that automatically expires after a preset period of time. The lease enables the lease owner exclusive use of the associated content in the memory system. Furthermore, the memory system has a memory access mechanism that is separate from the data users. The memory access mechanism includes hardware components to enforce lease-based access to the memory system. Since the lease expires rather than being permanently granted, no CPU or software thread is able to monopolize a lease forever. Moreover, since the lease-base access is hardware-assisted, faulty CPUs or software are prevented from unauthorized access to leased (or locked) content in which another CPU or software already holds a lease.

Abstract: A network and storage I/O device is described for use with a host computer system having a system bus coupled to a host processor and a main memory to provide a high bandwidth network server system. The network and storage I/O device includes a plurality of network controllers to communicate with client computers connected over a network, a plurality of storage controllers to transfer data to and from storage devices, at least one memory element to temporarily store data transferred between the network controllers and the storage controllers and a crossbar switch having a plurality of nodes to interconnect the plurality of network controllers, the plurality of storage controllers and the at least one memory element. The network and storage I/O device also includes a bridge coupled between one of the nodes and the system bus of the host computer.

Abstract: The present invention is directed to an apparatus and method to detect errors in bits of a binary coded data word transmitted from a sender device to a receiver device by having the sender device determine and store the parity of the data word, transmitting the data word to the receiver device, the receiver device determines the parity of the received data word, transmits the parity of the received data word to the sender device and the sender device compares the stored parity against the received parity. Variations include a pipelining of sending and receiving data words, using parity store registers to allow continuous cycles of transmission and only retransmitting if an error is detected. Another variation includes multiple devices as senders and receivers forming a virtual network in which sender and receiver devices can transmit data words and send back errors along the network.