Abstract: Data of a vector storage request pertaining to one or more disjoint, non-adjacent, and/or non-contiguous logical identifier ranges are stored contiguously within a log on a non-volatile storage medium. A request consolidation module modifies one or more sub-requests of the vector storage request in response to other, cached storage requests. Data of an atomic vector storage request may comprise persistent indicators, such as persistent metadata flags, to identify data pertaining to incomplete atomic storage requests. A restart recovery module identifies and excludes data of incomplete atomic operations.

Abstract: Apparatuses, systems, methods, and computer program products are disclosed for hybrid checkpointed memory. An extended memory module uses volatile memory of a host and a non-volatile memory medium as virtual memory for the host. A clone module clones data of a range of virtual memory in response to a checkpoint event for the range of virtual memory. A range of virtual memory may include data stored in a volatile memory and data stored in a non-volatile memory medium. A checkpoint module flushes dirty data of a range of virtual memory to a non-volatile memory medium in response to a checkpoint event. A hybrid checkpointed memory interface provides access to data of a range of virtual memory while dirty data is being flushed using data of a range of virtual memory, or using a clone of the data.

Abstract: Apparatuses, systems, program products, and method are disclosed for machine learning abstraction. An apparatus includes an objective module configured to receive an objective to be analyzed using machine learning. An apparatus includes a grouping module configured to select a logical grouping of one or more machine learning pipelines to analyze a received objective. An apparatus includes an adjustment module configured to dynamically adjust one or more machine learning settings for a logical grouping of one or more machine learning pipelines based on feedback generated in response to analyzing a received objective.

Abstract: A storage controller is configured to implement an atomic storage operation comprising a plurality of separate storage operations on a non-volatile storage medium. The storage controller may store persistent indicators to identify data that pertains to the atomic storage operation. An invalid shutdown may occur before the atomic storage operation is complete. A restart and recovery operation comprises a first scan of the non-volatile storage medium to identify data of the failed atomic storage operation. A physical trim note is stored on the non-volatile storage medium to identify the data of the failed atomic storage operation. The data may be identified by media address. Storage metadata is reconstructed in a second scan, which excludes the data and/or operations of the failed atomic storage operation.

Abstract: A storage layer of a non-volatile storage device may be configured to provide key-value storage services. Key conflicts may be resolved by modifying the logical interface of data stored on the non-volatile storage device. Resolving a key conflict may comprise identifying an alternative key and implementing one or more range move operations configured to bind the stored data to the alternative key. The move operations may be implemented without relocating the data on the non-volatile storage device.

Abstract: A data services module performs log storage operations in response to requests by storing data on one or more storage devices, and appending information pertaining to the requests to a separate metadata log. A log order of the metadata log may correspond to an order in which the requests were received, regardless of the order in which data of the requests are written to the storage devices. The requests may correspond to identifiers of a logical address space. The data services module implements an any-to-any translation layer configured to map identifiers of the logical address space to the stored data. The virtualization module may include a metadata management module configured to checkpoint the translation layer metadata by, inter alia, appending aggregate, checkpoint entries to the metadata log. The data services module may leverage the translation layer between the logical identifiers and underlying storage locations to efficiently implement logical manipulation operations.

Abstract: A cache module leverages storage metadata to cache data of a backing store on a non-volatile storage device. The cache module maintains access metadata pertaining to access characteristics of logical identifiers in the logical address space, including access characteristics of un-cached logical identifiers (e.g., logical identifiers associated with data that is not stored on the non-volatile storage device). The access metadata may be separate and/or distinct from the storage metadata. The cache module determines whether to admit data into the cache and/or evict data from the cache using the access metadata. A storage module may provide eviction candidates to the cache module. The cache module may select candidates for eviction. The storage module may leverage the eviction candidates to improve the performance of storage recovery and/or grooming operations.

Abstract: A storage module may be configured to perform log storage operations on a storage log maintained on a non-volatile storage medium. An I/O client may utilize storage services of the storage module to maintain an upper-level log. The storage module may be configured to coordinate log storage and/or management operations between the storage log and the upper-level log. The coordination may include adapting a segment size of the logs to reduce write amplification. The coordination may further include coordinating validity information between log layers, adapting log grooming operations to reduce storage recovery overhead, defragmenting upper-level log data within the storage address space, preventing fragmentation of upper-level log data, and so on. The storage module may coordinate log operations by use of log coordination messages communicated between log layers.

Abstract: A storage layer may over-provision physical storage resources of a storage medium by reserving a portion of the full physical storage capacity of the storage medium for use as reserve capacity. The reserve capacity may be used to prevent write stall conditions and/or for grooming operations, such as storage recovery, refresh, and the like. A reserve module may be configured to adapt the reserve capacity in accordance with, inter alia, operating conditions on the storage layer. The reserve module may be configured to dynamically modify the storage capacity available through the storage layer. A cache layer configured to cache data of a backing store on the storage layer, may be configured to add and/or remove cache entries in response to changes in the reserve capacity.

Abstract: An aggregation module combines a plurality of logical address spaces to form a conglomerated address space. The logical address spaces comprising the conglomerated address space may correspond to different respective storage modules and/or storage devices. An atomic aggregation module coordinates atomic storage operations within the conglomerated address space, and which span multiple storage modules. The aggregation module may identify the storage modules used to implement the atomic storage request, assign a sequence indicator to the atomic storage request, and issue atomic storage requests (sub-requests) to the storage modules. The storage modules may be configured to store a completion tag comprising the sequence indicator upon completing the sub-requests issued thereto. The aggregation module may identify incomplete atomic storage requests based on the completion information stored on the storage modules.

Abstract: A storage layer stores integrity data corresponding to data groups being written to a storage array. The storage layer validates stored data groups by use of the integrity data, independent of recovery information managed by the storage array. The storage layer preempts crash recovery operations of the storage array by validating a subset of the data groups. The storage layer identifies data groups to validate based on a location of a log append point within a storage address space of the storage array. The storage layer may be further configured to mitigate write hole conditions by, inter alia, reconstructing invalid data of a data group and/or retaining valid portions of the data group.

Abstract: A system and method facilitate processing atomic storage requests. The method includes receiving, from a storage client, an atomic storage request for a first storage device that is incapable of processing atomic write operations. The method also includes processing the atomic storage request at a translation interface. The method also includes storing the atomic storage request in one or more storage operations in a second storage device capable of processing the atomic storage request.

Abstract: A compression manager generates a reduced-size data set for a page. The reduced-size data set is stored on a storage medium by a storage module, and unused portions of the page are identified and/or recorded by the storage module. The reduced-size data set may be generated by compressing the page. The reduced-size data set may be written in response to a request from a client, which provides a message that identifies the unused portion(s) of the page. The page may be stored in a file, and the unused portion may correspond to empty portion(s) within the file. The storage module may maintain an index to map the page to the reduced-size data set, such that portions of the page are unmapped. The storage capacity used by the page may correspond to the mapped portion of the page, and exclude portion(s) of the page that are unmapped.

Abstract: An atomic storage module may be configured to implement atomic storage operation directed to a first set of identifiers in reference to a second, different set of identifiers. In response to completing the atomic storage operation, the atomic storage module may move the corresponding data to the first, target set of identifiers. The move operation may comprise modifying a logical interface of the data. The move operation may further include storing persistent metadata configured to bind the data to the first set of identifiers.

Abstract: Techniques are disclosed relating to enabling virtual machines to access data on a physical recording medium. In one embodiment, a computing system provides a logical address space for a storage device to an allocation agent that is executable to allocate the logical address space to a plurality of virtual machines having access to the storage device. In such an embodiment, the logical address space is larger than a physical address space of the storage device. The computing system may then process a storage request from one of the plurality of virtual machines. In some embodiments, the allocation agent is a hypervisor executing on the computing system. In some embodiments, the computing system tracks utilizations of the storage device by the plurality of virtual machines, and based on the utilizations, enforces a quality of service level associated with one or more of the plurality of virtual machines.

Abstract: Apparatuses, systems, methods, and computer program products are disclosed. A method includes tracking which portions of data stored in a volatile memory buffer are not yet stored in a non-volatile memory medium. A volatile memory buffer may be accessible using memory semantics. A volatile memory buffer may be associated with logic configured to ensure that the data stored in the volatile memory buffer is non-volatile. A method includes maintaining consistency of data between a volatile memory buffer and a non-volatile memory medium based on tracked portions of the data. A method includes copying at least portions of data not yet stored in a non-volatile memory medium to the non-volatile memory medium in response to a trigger.

Abstract: Apparatuses, systems, methods, and computer program products are disclosed for hybrid checkpointed memory. An extended memory module uses volatile memory of a host and a non-volatile memory medium as virtual memory for the host. A clone module clones data of a range of virtual memory in response to a checkpoint event for the range of virtual memory. A range of virtual memory may include data stored in a volatile memory and data stored in a non-volatile memory medium. A checkpoint module flushes dirty data of a range of virtual memory to a non-volatile memory medium in response to a checkpoint event. A hybrid checkpointed memory interface provides access to data of a range of virtual memory while dirty data is being flushed using data of a range of virtual memory, or using a clone of the data.

Abstract: A cache and/or storage module may be configured to reduce write amplification in a cache storage. Cache layer write amplification (CLWA) may occur due to an over-permissive admission policy. The cache module may be configured to reduce CLWA by configuring admission policies to avoid unnecessary writes. Admission policies may be predicated on access and/or sequentiality metrics. Flash layer write amplification (FLWA) may arise due to the write-once properties of the storage medium. FLWA may be reduced by delegating cache eviction functionality to the underlying storage layer. The cache and storage layers may be configured to communicate coordination information, which may be leveraged to improve the performance of cache and/or storage operations.

Abstract: Apparatuses, systems, methods, and computer program products are disclosed for providing a memory device with volatile and non-volatile media. A volatile memory medium is on a circuit board configured to be installed on a memory bus of a processor. A non-volatile memory medium is on the same circuit board. A mapping module is configured to selectively store data in either the volatile memory medium or the non-volatile memory medium. The data is provided by way of one or more commands from the processor.

Abstract: Techniques are disclosed relating to enabling virtual machines to access data on a physical recording medium. In one embodiment, a computing system provides a logical address space for a storage device to an allocation agent that is executable to allocate the logical address space to a plurality of virtual machines having access to the storage device. In such an embodiment, the logical address space is larger than a physical address space of the storage device. The computing system may then process a storage request from one of the plurality of virtual machines. In some embodiments, the allocation agent is a hypervisor executing on the computing system. In some embodiments, the computing system tracks utilizations of the storage device by the plurality of virtual machines, and based on the utilizations, enforces a quality of service level associated with one or more of the plurality of virtual machines.