International Business Machines Corporation

Vinodh Venkatesan

Ilias Iliadis

Giovanni Cherubini

Jens Jelitto

Slavisa Sarafijanovic

Yusik Kim

Mark Alfred Lantz

Abu Sebastian

Robert Haas

Robert Beverley Basham

Elliot Karl Kolodner

Dmitry Sotnikov

Paul Merrill Greco

Paula Kim Ta-Shma

Timoleon Moraitis

Taras Lehinevych

A computer-implemented method for managing datasets of a storage system is provided, wherein the datasets have respective sets of metadata, the method including: successively feeding first sets of metadata to a spiking neural network (SNN), the first sets of metadata fed corresponding to datasets of the storage system that are labeled with respect to classes they belong to, so as to be associated with class labels, for the SNN to learn representations of said classes in terms of connection weights that weight the metadata fed; successively feeding second sets of metadata to the SNN, the second sets of metadata corresponding to unlabeled datasets of the storage system, for the SNN to infer class labels for the unlabeled datasets, based on the second sets of metadata fed and the representations learned; and managing datasets in the storage system, based on class labels of the datasets, these including the inferred class labels.

Managing datasets of a cognitive storage system with a spiking neural network

A storage system and a method for storing a data segment, a storage capacity manager and a method for managing a capacity of a storage unit, and a storage tier relocation manager and a method for relocating a data segment. The storage system includes at least two storage tiers, an access pattern evaluator, a classification unit, a selector, and logic. The storage capacitor manager includes a monitoring unit and capacity managing unit. The storage tier relocation manager includes a target storage tier, the data segment relocated to the target storage tier, and a protection measure.

Automatically organizing storage system

A method for data storage tape access optimization is provided. The method identifies a plurality of data segments stored on a data storage tape. Each segment having a location and a size on the data storage tape. The method determines a subset of data segments stored on the data storage tape selected for retrieval. The method determines an access order for the subset of data segments. The access order represents a minimized tape cycle time for retrieving the subset of data segments. The method retrieves the subset of data segments from the data storage tape based on the access order.

Access ordering for tape cycle optimization

The present invention is notably directed to a computer-implemented method of accessing data by an application. The method comprises detecting an application that is accessing an initial file stored in a storage system; fetching an application model of the detected application, the application model comprising an association between access patterns of the detected application when accessing first files with first metadata information of the said first files; identifying one or more second files stored in the storage system, the identified one or more second files having second metadata information that are similar to the first metadata information; and prefetching the identified one or more second files.

Data prefetching for large data systems

Predictively selecting subset of disks of a storage system to be spun-up, including providing metadata of data entities stored in the disks of the storage system, estimating the data entity access probabilities for a prediction time window based on the metadata, each data entity access probability being indicative for the probability of access to a certain data entity within the prediction time window, calculating disk access probabilities for a prediction time window based on the estimated probability of access of data entities, each disk access probability being indicative for the probability of access to a certain disk within the prediction time window, estimating the number of disks to be spun-up in a certain prediction time window, dynamically adapting the data entity threshold value and/or the disk access threshold value, selecting a subset of disks to be spun-up in the following prediction time window.

Controlling a storage system

A computer-implemented method for managing a tiered storage system having an archive tier and an active storage tier comprises determining a workload for moving data between the active tier and the archive tier; and determining an assignment of data to be stored across the active tier and the archive tier, based on the determined workload.

Data archive tier of a tiered storage system

A computer-implemented method is provided for managing digital datasets stored on a multi-tiered storage system that includes several tiers of storage, the datasets likely to be accessed by one or more applications interacting with the storage system. The method includes monitoring an access history of datasets accessed by the one or more applications; and while monitoring the access history: computing probabilities of access, by the one or more applications, of the datasets stored on the storage system according to metadata associated to given datasets as identified in the monitored access history; and based on the computed probabilities of access, selecting one or more of the datasets to be moved across the tiers. Related storage systems and computer program products are also provided.

Managing digital datasets on a multi-tiered storage system based on predictive caching

The invention is notably directed to computer-implemented methods and systems for managing datasets in a storage system. In such systems, it is assumed that a (typically small) subset of datasets are labeled with respect to their relevance, so as to be associated with respective relevance values. Essentially, the present methods determine, for each unlabeled dataset of the datasets, a respective probability distribution over a set of relevance values. From this probability distribution, a corresponding relevance value can be obtained. This probability distribution is computed based on distances (or similarities), in terms of metadata values, between said each unlabeled dataset and the labeled datasets. Based on their associated relevance values, datasets can then be efficiently managed in a storage system.

Dataset relevance estimation in storage systems

An embodiment is directed to a method for determining an assignment of data to be stored on at least one storage tier i of a plurality of storage tiers of a tiered storage system. The method including, for the at least one storage tier i, steps of accessing storage device characteristics of the at least one storage tier i of the plurality; based on the accessed storage device characteristics, splitting the at least one storage tier i into Nstorage subtiers of the at least one storage tier i, the Nstorage subtiers having respective storage device characteristics; and based on characteristics of data to be stored on the tiered storage system and the respective storage device characteristics of the Nstorage subtiers, determining an assignment of data to be stored on each of the Nstorage subtiers. Embodiments are directed to related methods, systems and computer program products.

Subtier-level data assignment in a tiered storage system

A method for storing a data segment in a storage tier of a storage unit comprising at least two storage tiers includes receiving the data segment to be stored including metadata; receiving metadata of data segments stored in the storage unit; and determining the storage tier to store the received data segment to and a protection level dependent on the metadata received and dependent on the metadata of the received data segment.

Rule-based tiered storage system

A device and method for determining number of storage devices for each of plurality of storage tiers and assignment of data to be stored in the plurality of storage tiers. The device computes an optimized number of storage devices for each tier including a classifier, which receives units as input data to be stored in the system. Based on the characteristics of the storage devices, the device outputs a data-unit-to-storage tier assignment. An optimizer receives data-unit-to-storage tier assignment from the classifier and a cost budget for the system. Thereafter, an output for the number of storage devices for each storage tier is calculated. A method for determining number of storage devices for each of a plurality of storage tiers and assignment of data to be stored in the plurality of storage tiers are also provided.

Device and method for determining a number of storage devices for each of a plurality of storage tiers and an assignment of data to be stored in the plurality of storage tiers

Embodiments include methods for operating a first storage system having a first number of data storage drives for enabling access to a first set of removable media. Aspects include providing a second storage system having a number K of data storage drives for enabling access to a second set of removable media and providing a set of parameters describing operational characteristics of the second storage system. Aspects also include determining an analytical model using the set of parameters, the analytical model describing the variation of average waiting time as a function of system load over a predefined range covering multiple system load regime domains and determining values of the set of parameters using the analytical model and data of the second storage system. Aspects further include using the analytical model and the values of the set of parameters for reconfiguring the first storage system.

Method for dimensioning and reconfiguring of a storage system with removable media

A storage system is proposed which comprises a data storage device including a number N of removable storage media for storing data, with N&#x2267;2, and a number m of drives, with m&#x2267;1, wherein each of the m drives is configured to drive one of the removable storage media mounted to the drive. The storage system includes a server being coupled to the data storage device and configured to serve requests from clients, and a controller which is configured to control the data storage device to adaptively unmount the removable storage media based on a monitoring information of the requests at the data storage device, states of the drives, and states of the removable storage media.

Adaptively mounting and unmounting removable storage media based on monitoring requests and states of storage drives and the storage media

A method is provided for increasing data storage reliability in a heterogeneous storage system including multiple storage devices of different types. The devices store respective data subsets of a dataset. The method includes accessing configuration parameters for the dataset including first and second amounts of the data subsets respectively stored on the multiple storage devices. The method further includes estimating an initial global reliability of the heterogeneous storage system for the dataset, based on the configuration parameters. The method also includes determining an increased global reliability of the storage system for the dataset, by estimating a global reliability as would be obtained by modifying at least some of the configuration parameters, whereby at least the first and second amounts of the data subsets respectively stored on the devices are modified. The method further includes reconfiguring the dataset stored across the heterogeneous storage system according to the modified configuration parameters.

Growing community of inventors

Waedenswil, Switzerland

Vinodh Venkatesan