Abstract: In one embodiment, a method includes receiving a stored copy request indicating data and including a request identifier, scheduling publication of the data to an object store, and sending a subscription request indicating the data and the request identifier to the object store. In another embodiment, a method includes receiving a publication request indicating data to be copied, wherein the publication request is a recording request indicating a channel and a timespan, receiving one or more subscription requests for the data, receiving the data, and generating a copy of the data for each subscription request.

Abstract: Dynamic storage provisioning for nested clusters is disclosed. A hosted cluster (HC) storage provisioner, executing in a hosted cluster that is hosted by an infrastructure cluster (IC), receives a request to dynamically provision a hosted cluster (HC) persistent volume object that is coupled to a physical storage. The HC storage provisioner causes an IC control plane executing on the IC to generate IC volume metadata that is backed by a storage volume on the physical storage. The HC storage provisioner determines that the IC volume metadata has been generated. The HC storage provisioner creates HC volume metadata on the hosted cluster that is linked to the IC volume metadata, the HC volume metadata comprising an HC persistent volume object that represents a persistent volume for use by the hosted cluster that is backed, via the IC volume metadata, by the physical storage.

Abstract: Methods, apparatuses, and systems for tensor memory access are described. Multiple data located in different physical addresses of memory may be concurrently read or written by, for example, employing various processing patterns of tensor or matrix related computations. A memory controller, which may comprise a data address generator, may be configured to generate a sequence of memory addresses for a memory access operation based on a starting address and a dimension of a tensor or matrix. At least one dimension of a tensor or matrix may correspond to a row, a column, a diagonal, a determinant, or an Nth dimension of the tensor or matrix. The memory controller may also comprise a buffer configured to read and write the data generated from or according to a sequence of memory of addresses.

Abstract: According to certain aspects, a system includes a client device that includes a virtual machine (VM) executed by a hypervisor, a driver located within the hypervisor, and a data agent. The VM may include a virtual hard disk file and a change block bitmap file. According to some embodiments the driver intercepts write operations generated by the VM to store data in a sector, determines an identity of the sector based on the intercepted write operation, determines an entry in the change block bitmap file that corresponds with the first sector, and modifies the entry in the change block bitmap file to indicate that data in that sector has changed. The data agent may generate an incremental backup of the VM based on the change block bitmap file in response to an instruction from a storage manager, where the incremental backup includes the data in the sector where data was modified.

Abstract: A method for releasing memory allocated by a contiguous memory allocator that merges a to-be-released memory page with an adjacent free page to form a memory block that can be released more efficiently than would be the case when releasing a series of un-merged memory pages.

Abstract: Systems, apparatuses, and methods related to channel architecture for memory devices are described. Various applications can access data from a memory device via a plurality of channels. The channels can be selectively enabled or disabled based on the behavior of the applications. For instance, an apparatus in the form of a memory system can include an interface coupled to a controller and a plurality of channels. The controller can be configured to determine an aggregate amount of bandwidth used by a plurality of applications accessing data from a memory device coupled to the controller via the plurality of channels and disable one or more channels of the plurality of channels based, at least in part, on the aggregate amount of bandwidth used by the plurality of applications.

Abstract: Systems and methods for durable storage of storage volume “snapshots” are provided. Snapshots are stored as collections of snapshot data objects. To improve the durability of snapshot storage, physical deletion of snapshot data objects may be delayed for a period of time after the snapshot data objects are marked for deletion. Lists of the stored snapshot data objects and the snapshot data objects that make up active snapshots may be periodically analyzed. If there are any snapshot data objects that are part of active snapshots and are not present in the list of stored snapshot data objects, the snapshot data objects may be recovered before they are physically deleted.

Abstract: An apparatus comprises at least one processing device comprising a processor and a memory. The at least one processing device is configured to provision a first logical device on a first storage system, to configure the first logical device to support replication from the first storage system to a second storage system, and to initiate generation of a remote backup copy of a second logical device on the second storage system. In conjunction with performance of a verification process for the remote backup copy of the second logical device, the at least one processing device is configured to cause the first logical device to be paired with the remote backup copy of the second logical device, and to mount the first logical device so as to permit performance of the verification process for the remote backup copy utilizing data obtained from the remote backup copy via the first logical device.

Abstract: Embodiments are directed to providing a secure address translation service. An embodiment of a system includes DRAM for storage of data, an IOMMU coupled to the DRAM, and a host-to-device link to couple the IOMMU with one or more devices and to operate as a translation agent on behalf of one or more devices in connection with memory operations relating to the DRAM, including receiving a translated request from a discrete device via the host-to-device link specifying a memory operation and a physical address within the DRAM pertaining to the memory operation, determining page access permissions assigned to a context of the discrete device for a physical page of the DRAM within which the physical address resides, allowing the memory operation to proceed when the page access permissions permit the memory operation, and blocking the memory operation when the page access permissions do not permit the memory operation.

Abstract: Techniques for in-band locking of extents containing multiple pages of shared non-volatile data storage are disclosed in which inter-node lock request messages indicate both an individual page for which a lock is requested and the multi-page extent that contains that page. A page lock or an extent lock is granted to the requesting node based on the time since the last access to the extent by the node to which the request was sent. A generation number may be maintained in each node and stored in the per-extent lock table entries and per-page lock table entries such that pages accessed by a node within an extent while the extent is in extent-locked access mode may subsequently be accessed by the node during a page-locked access mode for the extent occurring after the extent-locked access mode ends based on page locks created by the node during the extent-locked access mode.

Abstract: One embodiment provides a system which facilitates organization of data. During operation, the system receives data associated with a logical block address (LBA) to be written to a non-volatile memory. The system stores, in a data structure, a mapping of a first physical block address (PBA) corresponding to the LBA to a first status for the data, wherein the first status indicates data validity and recovery being enabled for the data. Responsive to receiving a command to delete the data, the system modifies the first status to indicate data invalidity and recovery being enabled for the data. Responsive to receiving a command to recover the previously deleted data, the system modifies the first status to indicate data validity and recovery being enabled for the data.

Abstract: Technology for proactively allocating data storage resources to a storage object. A rate at which host I/O operations directed to the storage object are received and/or processed is monitored during a monitored time period, and a high activity time range is identified. An anticipatory time range is defined that is a range of time immediately preceding the high activity time range. During the anticipatory time range within a subsequent time period following the monitored time period, high performance non-volatile storage is allocated to the storage object that is available for processing host I/O operations directed to the storage object at the beginning of and throughout the high activity time range. A low activity time range may also be identified, and lower performance non-volatile storage may be allocated to the storage object within an anticipatory time range immediately preceding the low activity time range.

Abstract: A method of storing and deleting data maintains a record of backups of the stored data, the record defining an entry for each backup comprising details of the data that has been backed up and the location of the data that has been backed up. The record of backups of the stored data is updated with a new entry every time at least part of the stored data is stored at a new location different from the first location. Once an instruction is received to delete the stored data, the maintained record of backups of the stored data is accessed, and an instruction is transmitted for each respective entry in the record of backups, the instruction comprising an instruction to delete the stored data at the respective location for the respective entry.

Abstract: Example embodiments relate generally to data resynchronization methods and systems in continuous data protection (CDP) and more specifically to an input and output (I/O) filtering framework and log management system to seek a near-zero recovery point objective (RPO).

Abstract: Provided are a computer program product, system, and method for providing track format information when mirroring updated tracks from a primary storage system to a secondary storage system. The primary storage system determines a track to mirror to the secondary storage system and determines whether there is track format information for the track to mirror. The track format information indicates a format and layout of data in the track, indicated in track metadata for the track. The primary storage system sends the track format information to the secondary storage system, in response to determining there is the track format information and mirrors the track to mirror to the secondary storage system. The secondary storage system uses the track format information for the track in the secondary cache when processing a read or write request to the mirrored track.

Abstract: A system includes a memory including a ring buffer having a plurality of slots, a processor in communication with the memory, a guest operating system, and a hypervisor. The hypervisor is configured to detect a request associated with a memory entry, retrieve up to a predetermined quantity of memory entries in the ring buffer from an original slot to an end slot, and test a respective descriptor of each successive slot from the original slot through the end slot while the respective descriptor of each successive slot in the ring buffer remains unchanged. Additionally, the hypervisor is configured to execute the request associated with the memory entries and respective valid descriptors. The hypervisor is also configured to walk the ring buffer backwards from the end slot to the original slot while clearing the valid descriptors.

Abstract: In one example, a method for writing data includes receiving a write request and performing a first type of logging process in connection with the write request, and creating a corresponding first logging record. Additionally, a second type of logging process is performed in connection with the write request, and a corresponding second logging record created, where the second type of logging process is different from the first type of logging process. Next, a determination is made, as between the two logging records, which of the logging records requires the least amount of non-volatile random access memory (NVRAM), and the logging record that requires the least amount of NVRAM is written to the NVRAM.

Abstract: Tile based interleaving and de-interleaving of row-column interleaved data is described. In one example, the de-interleaving is divided into two memory transfer stages, the first from an on-chip memory to a DRAM and the second from the DRAM to an on-chip memory. Each stage operates on part of a row-column interleaved block of data and re-orders the data items, such that the output of the second stage comprises de-interleaved data. In the first stage, data items are read from the on-chip memory according to a non-linear sequence of memory read addresses and written to the DRAM. In the second stage, data items are read from the DRAM according to bursts of linear address sequences which make efficient use of the DRAM interface and written back to on-chip memory according to a non-linear sequence of memory write addresses.

Abstract: A pinned memory space for caching data can be provided in a data node. The data that is cached in the pinned memory space can be prevented from being swapped out. A virtual address can be assigned to the data. The virtual address can be mapped to a memory address of the data in the pinned memory space for accessing the data by an application. A first command can be received from the application for caching the data. The first command can indicate an attribute associated with the caching of the data. Responsive to receiving the first command from the application for caching the data, the data associated with the first command can be cached by storing the attribute in association with the data in the pinned memory space.

Abstract: Embodiments of the disclosure provide a computer-implemented method for data replication from a first storage to a second storage. In the method, data replication control blocks for data blocks stored in the first storage and needed to be transmitted may be generated. The data replication control blocks may be sorted in an ascending order of probabilities of expected overwrites of the respective data blocks. The data blocks may be transmitted from the first storage to the second storage according to the sorted data replication control blocks.