KR20190088874A - A system and method for hybrid data reliability for object storage devices - Google Patents

Abstract

According to an embodiment of the present invention, a method for storing data in a key-value reliability system including the N number of storage devices grouped into a reliability group as a single logical unit and managed by a virtual device management layer comprises the steps of: determining whether data meets a threshold corresponding to a reliability mechanism for storing the data; selecting the reliability mechanism if the threshold is met; and storing the data according to the selected reliability mechanism, wherein N is an integer.

Description

[0001] SYSTEM AND METHOD FOR HYBRID DATA RELIABILITY FOR OBJECT STORAGE DEVICES [0002]

One or more aspects of embodiments of the present invention generally relate to data storage systems, and more particularly, to a key-value reliability system that includes a plurality of key- And a method for selecting a reliability mechanism for storing the data.

Data reliability mechanisms, such as erase coding, can be used to address data loss due to storage device failures and data corruptions in a variety of facilities including a plurality of storage devices .

Conventional solid state drives (SSDs) generally use only a block interface, erasure coding, or replication through a redundant array of independent disks (RAID) Lt; RTI ID = 0.0 > reliability. ≪ / RTI > As object formats vary in size and are not organized, efficient data conversion is required between object and block level interfaces. Furthermore, there is a need to ensure data reliability while maintaining the communicative efficiency and fast access time characteristics.

Techniques such as RAID are well studied for conventional block storage devices. However, the relatively new key-value storage devices may have conventional block devices and other interfaces and other storage semantics. Thus, a variety of new key-value storage devices may have potential benefits from new data reliability techniques that fit or are tailored to key-value data and key-value storage devices.

Since the reliability mechanisms of the embodiments can each perform a single key recovery procedure that allows the virtual device management layer to recover all keys present in the failed memory device and copy them to the new memory device, And can provide improvements in storage.

According to an embodiment of the present invention, there is provided a data storage method of a key-value reliability system including N (where N is an integer) storage devices grouped into a reliability group as a single logical unit and managed by a virtual device management layer / RTI > The method comprising the steps of: determining if the data satisfies a threshold corresponding to a reliability mechanism for storing the data; if the threshold is satisfied, selecting the reliability mechanism; .

Wherein the threshold is one or more of an object size of the data, a throughput consideration of the data, a read / write temperature of the data, and a basic erasure coding capability of the N storage devices .

The method further includes testing the data for the reliability mechanism using one or more Bloom filters or caches.

The method comprising: selecting the reliability mechanism, one or more checksums for each of the N storage devices storing the data, an object size of the data values stored in each of the N storage devices storing the data, And inserting metadata for recording the location of the parity group members of the N storage devices indicating which of the N storage devices have been stored with a key corresponding to the data .

Wherein the selected trust mechanism comprises an object copy, the step of storing the data comprises selecting a KV value, computing a hash for hashing the key corresponding to the selected KV value, Identifying some of the N storage devices to store replicas of objects, and writing the updated values corresponding to the KV value to each of the identified storage devices under the same user key name .

Wherein the selected reliability mechanism comprises packing and wherein storing the data comprises storing k key objects stored in k (where k is an integer) storage devices of the N storage devices of the trust group The method comprising: retrieving k number of value objects corresponding to the k number of key objects, selecting one of the k number of value objects having the largest value size among the k number of value objects so that the virtual sizes of all k number of value objects become equal (K) key objects to the k storage devices, the method comprising the steps of: (a) padding virtual zeroes to endpoints of the value objects, generating r (where r is an integer) parity objects from the k key objects, And writing the r parity objects to r storage devices among the N storage devices, Each of the support device is separated from the k number of storage devices, a single, k + r = N.

The selected reliability mechanism includes packing using typical erasure coding, and the N storage devices are configured with typical (k, r) maximum distance separable (MDS) erasure coding.

The selected reliability mechanism includes packing using regenerative erasure coding, and the N storage devices are configured with (k, r, d) regenerative erasure coding.

Wherein the selected reliability mechanism includes splitting, wherein storing the data comprises: selecting a KV value; dividing the KV value into k (where k is an integer) equal sized objects; Generating a parity object of r (where r is an integer) number of objects of the same size, computing a hash for hashing the key corresponding to the selected KV value, Determining the major device in which the KV value will be located, and determining, for each of the K storage devices, the k identical-sized objects and the r parity objects in the consecutive order starting from the main device to the N storage devices , Where k + r = N.

The selected reliability mechanism includes splitting using typical erasure coding, and the N storage devices are configured with typical (k, r) maximum distance separable (MDS) erasure coding.

Wherein the selected reliability mechanism comprises packing using regenerative erasure coding, wherein the N storage devices are configured with (k, r, d) regenerative erasure coding, and wherein the step of storing the data comprises using the regenerative erasure coding Dividing the k equal sized objects into m (where m is an integer) subpackets, and dividing each of the r parity objects into m parity subpackets.

According to another embodiment of the present invention, a data reliability system for storing data based on a selected reliability mechanism is provided. The data reliability system comprising: N (where N is an integer) storage devices configured as virtual devices using stateless data protection; And a virtual device management layer configured to manage the N storage devices as the virtual device and store data in selected ones of the N storage devices according to a mechanism, Determining whether the threshold meets a threshold corresponding to a reliability mechanism for storing the data, selecting the reliability mechanism if the threshold is satisfied, and storing the data according to the selected reliability mechanism.

Wherein the selected trust mechanism comprises an object copy, the virtual device management layer selects a KV value, computes a hash for hashing the key corresponding to the selected KV value, and replicates a key object corresponding to the KV value And storing the data by writing updated values corresponding to the KV value to each of the identified storage devices under the same user key name to identify the storage devices of the N storage devices .

Wherein the selected reliability mechanism comprises packing and the virtual device management layer selects k key objects stored in k (where k is an integer) storage devices among the N storage devices, and the k keys Wherein the k value objects corresponding to the objects are collected and the virtual value magnitudes of all of the k value objects are equal, (R, integer) parity objects from the k key objects, writes the k key objects to the k storage devices, and maps the r parity objects to the k storage devices Wherein each of the r storage devices is configured to store the data by writing to r storage devices of the N storage devices, Of being distinguished from the storage device, and r + k = N.

Wherein the selected trust mechanism comprises splitting, the virtual device management layer chooses a KV value, divides the KV value into k (where k is an integer) equal sized objects, , Wherein r is an integer, calculates a hash for hashing the key corresponding to the selected KV value, and calculates the KV value of the N storage devices based on the hash, And writing each of the k identical-sized objects and the r parity objects to the N storage devices in a sequential order starting from the main device, thereby storing the data , Where k + r = N.

Wherein the selected reliability mechanism comprises splitting using regenerative erasure coding, the N storage devices are configured with (k, r, d) regenerative erasure coding, and the virtual device management layer uses the regenerative erasure coding And is further configured to store the data by dividing the k pieces of the same sized objects into m (where m is an integer) subpackets and dividing each of the r parity objects into m parity subpackets .

According to another embodiment of the invention, when executed by a processor, storing data in a key-value reliability system comprising N storage devices grouped into a reliability group as a single logical unit and managed by a virtual device management layer There is provided a non-transient computer readable medium comprising computer code in which a method is performed. The method comprising the steps of: determining if the data satisfies a threshold corresponding to a reliability mechanism for storing the data; if the threshold is met, selecting the reliability mechanism; and storing the data according to the selected reliability mechanism .

Wherein the selected trust mechanism comprises an object copy, the step of storing the data comprises selecting a KV value, computing a hash for hashing the key corresponding to the selected KV value, calculating a key corresponding to the KV value Selecting some of the N storage devices to store a copy of the objects, and, under the same user key name, writing updated values corresponding to the KV value to each of the determined some storage devices .

Wherein the selected reliability mechanism comprises packing and wherein the storing the data comprises selecting k key objects stored in k (where k is an integer) storage devices of the N storage devices of the reliability group , Retrieving k value objects corresponding to the k key objects, calculating a value object having the largest value size among the k value objects so that the virtual value magnitudes of all the k value objects become equal, Generating r (where r is an integer) parity objects from the k key objects, writing the k key objects to the k storage devices, And writing the r parity objects to r storage devices of the N storage devices, wherein the r stories Each of the storage devices is distinguished from the k storage devices, and k + r = N.

Wherein the selected reliability mechanism includes splitting, wherein storing the data comprises: selecting a KV value; dividing the KV value into k (where k is an integer) equal sized objects; Generating r (where r is an integer) parity objects from the same sized objects, computing a hash for hashing the key corresponding to the selected KV value, Selecting one of the storage devices to which the KV value is to be placed; and selecting, from the main device, the k number of identical sized objects and each of the r parity objects in the N storage Lt; RTI ID = 0.0 > k + r = N. ≪ / RTI >

The foregoing and / or other aspects of the invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings.
1 is a block diagram illustrating a key-value reliability system for storing key-value data based on a selected trust mechanism, in accordance with an embodiment of the present invention.
Figure 2 is a flow diagram illustrating the selection of a reliability mechanism to be used by a key-value reliability system based on a size threshold corresponding to a size of data of a key-value pair, in accordance with an embodiment of the present invention.
Figure 3 is a block diagram of an embodiment of the present invention configured to store key-value data according to a trust mechanism of K-object (k, r) erasure coding, or multiple object "Packing" using typical erasure coding Lt; / RTI > is a block diagram showing a group of KV storage devices.
Figure 4 illustrates the storage of value objects and parity objects according to a reliability mechanism of K-object (k, r) erasure coding, or multi-object "packing" using typical erasure coding, FIG.
Figure 5 is a block diagram of an exemplary embodiment of a KV (k, r) erasure coded or "Splitting" trust mechanism configured to store key-value data, using typical erasure coding, ≪ / RTI > is a block diagram illustrating a group of storage devices.
Figure 6 is a block diagram of an embodiment of the present invention configured to store key-value data according to a reliability mechanism of single object (k, r, d) erasure coding, or "Splitting" using regenerative erasure coding Lt; / RTI > is a block diagram showing a group of KV storage devices.

The various features of the present invention and methods of achieving the present invention can be understood in more detail with reference to the following detailed description of the accompanying drawings and embodiments. In the following, embodiments will be described in more detail with reference to the accompanying drawings. Like numbers refer to like elements throughout. It should be understood, however, that the invention may be embodied in various other forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided as illustrations to assist in an overall understanding of the present invention, and will fully convey the technical features and aspects of the present invention to those skilled in the art. Accordingly, the processes, elements, and techniques that are not required for those of ordinary skill in the art to fully understand the features and aspects of the present invention may not be described. Unless otherwise stated, throughout the accompanying drawings and the detailed description that follows, like reference numerals refer to like elements, and their description will not be repeated. In the drawings, the relative sizes of elements, layers, and regions may be exaggerated for clarity.

Various embodiments are described herein with reference to cross-sectional views that are schematic illustrations of embodiments and / or intermediate structures. As such, variations from the shapes shown, for example, as a result of manufacturing techniques and / or tolerances, should be expected. Moreover, the specific structural or functional description set forth in the text is merely illustrative of the embodiments in accordance with the teachings of the present invention. That is, the embodiments described herein should not be construed as limited to the specifically illustrated forms of the areas, but should include variations in shape, for example, as a result of enhancement. For example, implant regions shown in a rectangle will typically have round or curved features and / or slopes of implant concentration at their edges rather than binary-change from implant to non-implant region . Likewise, the buried region formed by the implant can cause some implants in the region between the buried region and the surface on which the implant dies. That is, the areas shown in the figures are schematic in nature, and their shapes are not intended to be limiting of, and not limiting, to the actual shape of the area of the device. In addition, as will be appreciated by those skilled in the art, the described embodiments may be modified in various other ways without departing from the spirit or scope of the invention.

In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the conveniences and various embodiments of the present invention. However, various embodiments may be implemented without detailed description or with one or more equivalent alternatives. In other instances, well-known structures and devices are shown in block diagram form in order not to unnecessarily obscure the various embodiments.

The terms "first," "second," "third," and the like are used interchangeably throughout this document to describe various elements, structures, regions, layers, and / But it will be understood that such elements, structures, regions, layers, and / or regions are not limited to these terms. These terms are used only to distinguish one element, structure, region, layer, or region from another element, structure, region, layer, or region. That is, the first element, structure, region, layer, or section described below may be referred to as a second element, structure, region, layer, or section without departing from the spirit and scope of the present invention.

Spatially relative terms, such as "beneath, below, lower, under", "above", and the like, refer to the other element (s) Or may be used in the text to readily describe the relationship between the feature (s) and one element or feature (s). It will be appreciated that spatially relative terms are intended to encompass different orientations of the device in operation or use, in addition to the orientation shown in the figures. For example, in the drawings, when an apparatus is inverted, elements described as "below or beneath or under" other elements or features will point to "above" other elements or features. That is, exemplary terms of "below, under" may include both up and down directions. The device may be oriented in a different direction (e. G., Rotated 90 degrees or in the other direction) and the spatially relative descriptions used in the text should be interpreted accordingly. Similarly, if the first portion is described as being aligned with the second portion "on ", it is understood that without limitation to its upper surface based on gravity direction, And the lower surface is aligned.

Layer, region, or configuration is referred to as being "on", "connected to", or "coupled to" another element, layer, directly connected, or may have one or more intermediate elements, layers, regions, or configurations. However, the term "directly connected" refers to the direct connection of one component to another without intermediate configuration. On the other hand, other expressions describing the relationship between configurations such as " between immediately immediately "or" adjacent to or directly adjacent to " Or between layers, it is to be understood that there may be only one element or layer between elements or configurations, or one or more intermediate elements or layers may be present.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. As used in this text, singular terms are intended to include the plural forms, unless the context clearly dictates otherwise. Where the term "comprises" is used in the detailed description, the term " comprising " when used in the detailed description defines the presence of stated features, integers, steps, operations, elements, and / Steps, operations, elements, configurations, and / or groups thereof. As used herein, the terms "and / or" include any and all combinations or portions of one or more of the lists listed in the associated context.

As used herein, the terms "substantiall", "about" and similar terms are used as approximate terms and are not used as terms of degree, Is intended to describe deviations inherent in the measured or calculated values that can be recognized by those skilled in the art. As used herein, the term "about, approximately" includes values mentioned, and is used by those skilled in the art, taking into account errors associated with a particular amount of measurement and suspicious measurements (e.g., Quot; means within an allowable deviation for a particular value determined. For example, "about" may mean within one or more standard deviations or within ± 30%, 20%, 10%, 5% of the stated value. As used herein, the terms "use, using, and used" may be considered synonymous with "utilize, utilized, and utilized". Also, the word "exemplary" is intended to refer to an "exemplary " or explanatory illustration.

If a particular embodiment is implemented differently, then the particular process order may be performed differently from the order described. For example, two consecutively described processes may be performed substantially concurrently or in the reverse order described.

The electrical and / or electronic devices and / or other associated devices or configurations in accordance with embodiments of the invention described herein may be implemented within a computer-readable medium such as a computer-readable medium, such as, for example, Software, or a combination of software, firmware, and hardware. For example, various configurations of such devices may be formed in an integrated circuit (IC) chip or in separate IC chips The various configurations of these devices may be implemented in a flexible printed circuit film, a tape carrier package (TCP), a printed circuit board (PCB) various configurations of such devices may be used to perform various functions described in the text, And may be a process or thread running on one or more processors in one or more computing devices that communicate with other system configurations and execute computer program instructions. The computer program instructions may be stored in other non-volatile computer readable media, such as, for example, a CD-ROM, flash drive, or the like. It should be understood by those skilled in the art that the functions of the various computing devices may be combined or integrated into a single computing device without departing from the spirit and aspects of the exemplary embodiments of the present invention, It will be understood that it may be distributed to other computing devices There.

Unless defined otherwise, all terms including technical / scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in the common dictionary shall be construed as having a meaning consistent with their meaning in the context of the related art and / or in the detailed description of the present invention, and are, unless defined specifically in the text, It should not be interpreted in an overly formal sense.

As will be described below, embodiments of the present invention provide a key-value reliability system comprising a plurality of key-value (KV) storage devices grouped into one logical unit, - Provides a way to reliably store value data. Furthermore, embodiments of the present invention provide a stateless high reliability manager that manages drives and controls the storage of key-value (KV) pairs. Stateless Hibris Reliability Manager supports object replication; K-object (k, r) erasure coding-packing (K-Object (k, r) erasure coding-packing); Single Object (k, r) Erasure Coding - Splitting (Single Object (k, r) Erasure Coding - Splitting); K-object (k, r, d) regeneration coding-packing (K-Object (k, r, d)); , And a plurality of pluggable reliability mechanisms / techniques / implementations including a single object (k, r, d) regeneration coding-splitting (Single Object (k, r, d) regeneration coding-splitting).

A stateless hybrid reliability manager that is dependent on a plurality of pluggable reliability mechanisms can manage devices, control storage of KV pairs, and methods described for the selection of reliability mechanisms can efficiently store KV pairs of different sizes , Recovery, and recovery, the described embodiments may improve memory storage (e.g., storage of key-value data in key-value storage devices).

1 is a block diagram illustrating a key-value reliability system for storing key-value data based on a selected trust mechanism, in accordance with an embodiment of the present invention.

1, a variety of new key-value (KV) storage devices / memory devices / drives / KV-SSDs 130 are coupled to key-value data and KV storage devices 130, Can be potentially benefited from the new data reliability mechanism employed or tailored to the system. Thus, a hybrid key-value reliability system for these KV storage devices 130 manages the KV storage devices 130 using a hybrid reliability mechanism in accordance with one or more pluggable reliability mechanisms < RTI ID = 0.0 > , And a stateless hybrid reliability manager / virtual device manager layer / virtual device management layer 120 that controls the storage of KV pairs in them. Although solid-state drives are commonly used to refer to the KV storage devices described herein, other storage devices may be used in accordance with embodiments of the present invention. The design and operation of the virtual device management layer 120 according to embodiments of the present invention will be described below.

In the current embodiment, virtual device management layer 120 may enable a method for reliably storing key-value data / KV pair 170 in a key-value trusted system. The key-value reliability system may include a plurality of KV storage devices 130 organized in one logical unit. The logical unit may be referred to as the reliability group 140.

The KV storage devices 130 of the trust group 140 may store respective chunks of erasure coded data and / or replicated data, which may correspond to the key-value data 170. The KV storage devices 130 of the trust group 140 appear as a single virtual device 110 in which key-value operations are directed through the virtual device management layer 120. [

The virtual device 110 may include a stateless hybrid reliability manager as the virtual device management layer 120. That is, the virtual device management layer 120 may operate in a stateless manner (i.e., it does not need to maintain the mapping between the key-value and the device).

Thus, the virtual device 110 may include N KV storage devices 130 (where N is an integer) (e.g., KV-SSDs 130-1, 130-2, 130-3, 130-4, ...). Value data 170 via the virtual device management layer 120 and store the key-value data 170 in the KV storage devices 130 via the virtual device management layer 120. The key- That is, the virtual device management layer 120 can manage the KV storage devices 130 and can control storing KV pairs in them.

In other embodiments, the key-value reliability system may also include a cache that selectively stores metadata and / or data associated with the keys of the key-value data 170 to improve the speed of operation. Reliability mechanisms may attach values of KV pairs to include metadata corresponding to KV pairs. That is, both the key and the value can be attached with the metadata identifier "MetaID" and the corresponding information to store additional metadata specific to the KV pair. The metadata may include a checksum, a reliability mechanism identifier to identify the reliability mechanisms used to store the data, an erasure code identifier, object sizes, location of parity group numbers, and so on.

In other embodiments, the key-value reliability system may further include bloom filters corresponding to the reliability mechanisms described below. Bloom filters can store the stored keys using a corresponding reliability mechanism, and thus can help the key-value reliability system in read operations. Thus, one or more Bloom filters or caches in a key-value reliability system may be able to quickly test keys against existing reliability mechanisms.

Each of the reliability mechanisms described herein uses key-value data 170 and a corresponding KV pair 170 using the same hash function for the key modulo of the corresponding one of the KV storage devices 130. [ You can save the first copy or chunk first. That is, for each of the pluggable reliability mechanisms, the trust mechanism may store at least the first copy / chunk using the same key as the user key.

As described above, embodiments of the present invention provide multiple pluggable reliability mechanisms to ensure reliable storage of key-value data 170 in a plurality of KV storage devices 130 do. Thus, the virtual device management layer 120 may require reliability mechanisms and may determine which of the reliability mechanisms is used.

The trust mechanisms may be based on policies such as value-size thresholds set during installation of virtual machine 110, and / or object read / write frequency. Thus, the virtual device management layer 120 may select an appropriate reliability mechanism based on the mentioned policies of the system.

For the five reliability mechanisms of the embodiments of the present invention, how the reliability mechanism operates by the virtual device management layer 120, and when the reliability mechanism can be used and selected appropriately is described below. These reliability mechanisms include Object Replication, K-Object (k, r) Erasure Coding-Packing (K-Object) (K, r, d) regeneration coding-packing), a single object (k, r) erasure coding-splitting, (k, r, d) regeneration coding (Splitting).

FIG. 2 is a flowchart 200 illustrating the selection of a reliability mechanism used by a key-value reliability system based on a magnitude threshold corresponding to a magnitude of KV pair of data, in accordance with an embodiment of the present invention.

Referring to Figure 2, for the entire supported reliability mechanisms (e.g., five of the above-described reliability mechanisms) based on size thresholds, the virtual device management layer 120 provides data (e.g., key- Data 170), determine whether the value of the value is less than a given threshold (t _i ) corresponding to the respective reliability mechanism, and determine a first reliability mechanism that meets the value-size threshold requirement You can choose.

For example, at S210, the virtual device management layer 120 may receive supported reliability mechanisms based on "n " (n is an integer) size thresholds. At S220, the virtual device management layer 120 may simply review each of the reliability mechanisms, one at a time, in the order of 1 to n. In S230, supported review of each of the reliability mechanism that is, the virtual device management layer 120 may have the size of the value data can be determined, it is smaller than the threshold value (t _i) for each of the reliability and response mechanism.

At S240, if a trust mechanism with a threshold (ti) greater than or equal to the value size of the data is found, the virtual machine management layer 120 may select its trust mechanism for use. In S250, if the reliability mechanism to be used in S240 is determined, or if it is determined in the final iteration of S220 that the reliability mechanism with the appropriate threshold (t _i ) satisfying the value size is not among the n reliability mechanisms, 120 may terminate the determination of the reliability mechanism to be used.

In the present embodiment, "n" may be equal to 5 according to five different reliability mechanisms of the embodiments described herein. For relatively small key-values (i. E., When the value size is relatively small), the virtual device management layer 120 may select a reliability mechanism for object replication for use. For slightly larger key-values, the virtual device management layer 120 may choose a packing mechanism, and then a splitting (e.g., in order) reliability mechanism, while each A typical erasure coding can be used. However, for larger key-values, virtual device management layer 120 may choose packing, then splitting, while using regeneration erasure coding instead of typical erasure coding, Can be used.

In embodiments of the present invention, the choice of the trust mechanism for use may be based on the object size of the object, the throughput requirements for the object, the read / write temperature of the corresponding key-value pair, And / or detection of whether the key is hot or cold. ≪ RTI ID = 0.0 > For example, "hot" keys can use the reliability mechanism of Object Replication, regardless of their value size, while "cold" Can be applied as one of the reliability mechanisms of the erasure coding scheme. As another example, the determination of whether to use the reliability mechanism of Object Replication may be based on both size and write temperature. Thus, a threshold corresponding to the object read / write period may be used to determine the reliability mechanism, instead of a threshold corresponding to magnitude in the flowchart 200 of FIG.

The respective operations of the five reliability mechanisms are described below.

Referring back to Figure 1, as previously mentioned, if the value size of the KV pair is relatively small, then the reliability mechanism of "Object Replication" is suitable for selection by the virtual device management layer 120 . Object Replication may be applied per object (e.g., per KV pair / key-value data 170). Although the reliability mechanism of Object Replication has high storage overhead, it has low read and recovery costs and can be suitable for very small value sizes.

The trust mechanism of object replication may also be suitable for key-values (e. G., Key-value data 170) with frequent updates, and thus may be selected based on read and write cycles .

During object replication, each time a write occurs, the object / key-value data 170 is replicated to one or more additional KV storage devices 130. A primary copy of the key-value data 170 may be located in one of the KV storage devices 130 designated by the hash of the key module N. Replicas of key copies of the key-value data 170 may be placed in a circular fashion, in successive KV storage devices 130, or immediately adjacent KV storage devices 130 have.

The virtual device management layer 120 or the user can determine how many copies of the key-value data 170 should be generated. For example, a distributed system using the virtual device management layer 120 may select 3-way replication and may basically create 3-way replication. However, users of the system can configure the number of replicas of the object to be more or less than the selected default value.

Thus, for example, if three-way replication is used and the primary KV storage device 130-2 includes a major copy of the data (e.g., key-value data 170), then the virtual device management layer 120 may store a replica of the primary copy of the data in subsequent replica KV storage devices 130-3 and 130-4, and all copies of the data are identical. That is, the copies of the data may include two (or more) subsequent KV storage devices 130-3, 130-4 (also referred to as " KV storage devices ") immediately following the KV storage device 130-2 For example, a circulation method).

Copies of the data may be stored in duplicate KV storage devices 130-3 and 130-4 under the same key name / same user key as the primary KV storage device 130-2. All copies of the data may include an identifier and a checksum to identify the replicated key-value data 170.

Thus, if a particular KV storage device 130 fails (e.g., KV storage device 130-3 fails), the KV storage devices 130 immediately before and after the failed KV storage device 130 By restoring the values using a recovery mechanism for the key names in the KV storage devices 130-2 and 130-4 immediately before and after the KV storage device 130-3, for example, Ensuring that the replicated key is recovered.

In summary, the virtual device management layer 120 can receive the key-value data 170 and hash the key object to create a KV storage device < RTI ID = 0.0 > (130). Virtual device management layer 120 may then update the updated values to selected KV storage devices 130 (e.g., selected KV storage devices (e. G. 130-2, 130-3, and 130-4.

FIG. 3 is a block diagram of an embodiment of a K-object (k, r) erasure coding or a multi-object "packing (K, r) erasure coding using traditional erasure coding, ≪ / RTI > is a block diagram illustrating a group of KV storage devices configured to store key-value data in accordance with a reliability mechanism of a " Packing "

With reference to FIG. 3, the reliability mechanism of packing using typical erasure coding is that the data having small value sizes that are not suitable for partitioning into chunks (e.g., for better data throughput) Can be selected. For example, the reliability mechanism of Packing using typical erasure coding may be based on the larger value sizes (but still relatively large) than the value sizes that are the result of the selection of the reliability mechanism of the Object Replication described earlier May be selected by the virtual device management layer 120 for data having a small size (e.g., small).

Packing using typical erasure coding can be composed of typical (k, r) maximum distance separable (MDS) erasure coding and can be used with systematic MDS codes. For example, the erase code may be essentially (4,2) Reed-Solomon code, (4,2) Reed-Solomon code is relatively well studied, and the corresponding fast implementation library Is readily available.

In using the reliability mechanism of packing using typical erasure coding, k keys / key objects from queues of k different KV storage devices 330 that are part of the same parity group / Are selected and erasure-coded and packed. (k is an integer).

For example, virtual device management layer 120 may store recently written key objects for each KV storage device 330 (e.g., each KV storage device 130 in the trust group 140 of FIG. 1) Maintains a buffer of the KV pairs 350 to allow the virtual device management layer 120 to select and cancel the k key objects 350 from k different KV storage devices, k key objects 350 are packed.

In the current embodiment, the virtual device management layer 120 includes four key objects 350x, 350y, and 350b, respectively, from four different KV storage devices 330-1, 330-3, 330-4, , And 350c. (In the present embodiment, k = 4)

Figure 4 illustrates the storage of value objects and parity objects according to a reliability mechanism of K-object (k, r) erasure coding, or multi-object "packing" using typical erasure coding, FIG.

3 and 4, key objects 350 are located in ^th KV storage device 330 (hash of n as key module). That is, for each key object 350, each hash of n into a key module that can be sent to a queue of a particular KV storage device 330 may be performed. In the present embodiment, the i-th key (Key _i ) 350-i is hashed and located in KV-SSD1 330-1, the j-th key (Key _j ) 350- -SSD2 330-2, and the k-th key (Key _k ) 350-k is hashed and located in the KV-SSD4 330-4.

The user value length / value size 462 of each stored value objects 450 is the same as written. However, for consistency in enabling erasure coding, the user value / value objects 450 have all the same size by having attached "0" peels / virtual zeroes / virtual zero padding 464 with them Respectively. That is, since each of the user value sizes 462 of the other value objects 450 may be changed (i.e., the value objects 450 may have distinct lengths or may be fractional-length key values). ), By implementing the method of virtual zero padding 464 (i. E. By padding the zeroes of virtual zero padding 464 with value objects 450 for coding, while implementing value objects 450 containing padded zeros 450 The parity objects 460 may have the same size as the largest object (s) 470 in the parity group 340, Thus, in the current embodiment, the value objects 450 "Val x", "Val y", and "Val b" are padded with virtual zeros that are not actually stored in the KV storage devices, 470) "Val c ". Thus, parity objects 460 can be computed.

After coding the k key objects 350, the virtual device management layer 120 extracts r parity objects from k key objects 350 and corresponding k values / k value objects 450, (460). Where k is an integer and k + r = N and N is the KV storage devices 330 of the parity group 340 (e.g., N KV storage devices of the reliability group 140 of FIG. 1) (130).

The virtual device management layer 120 allocates r parity objects 460 to r other remaining KV storage devices 330 of the parity group 340 (i.e., k key objects 350 are selected and erasure- Lt; RTI ID = 0.0 > kV < / RTI > Each of the k key objects 350 and the r parity objects 460 may be stored in each of the N storage devices 330 and the corresponding data may be stored in the parity group 340 Is evenly distributed to each of the N KV storage devices 330. [

Although readings and writes are relatively simple relative to the reliability mechanism of packing using typical erasure coding, parity re-arithmetic and recovery may be less simple. In order to recognize which key objects 350 are grouped together in the same parity group 340 to thereby perform the operation of the parity for the restoration and operation of the parity (for example, in the case of an update) The information about the groups of objects 350 is stored in the KV storage device 450 together with the actual value size 462 for each value object 450 (i.e., the value size 462 without the virtual zero padding 464) (E.g., KV storage devices 130 of FIG. 1). Thus, in the current embodiment, additional metadata is associated with key objects 350 (e.g., key objects 350 located in the trust group 140 of FIG. 1), key objects 350 corresponding to key objects 350 May be used to store KV storage devices 130 in the original length of each of the value objects 450, and in the coding order of the key objects 350.

For example, the metadata object value may refer to the entire key objects 350 of the trust group 140 and may also include a field indicating the value sizes 462 of the value objects 450, (I.e., value objects 450 containing the zeros of the virtual zero padding 464), value sizes 462 of the parity objects 460, and r parity objects 460 stored Lt; RTI ID = 0.0 > KV < / RTI >

The data may be stored using a user key. The metadata may be stored in an internal key formed using a MetaID indicator denoting the user key and "Metadata ". Moreover, the value sizes 462 may be used for accurate reconstruction when the value objects 450 are regenerated by determining where the value objects 450 terminate and where the zeroes of the virtual zero padding 464 begin And may be stored in the meta data to recognize the location of the virtual zero padding 464 (i.e., where zeros are added).

If one of the KV storage devices 330 fails, both data and metadata may be stored in the same KV storage device 330, so that data and metadata may potentially be lost, It may not be possible. However, in order to prevent such a situation, the metadata object value is set to "Object Replication Engine (Object) " of the virtual device management layer 120, which can implement the reliability mechanism of the previously mentioned Object Replication for the metadata object value Replication Engine) ".

In addition, if the KV storage device 330 supports object linking, the same metadata object values may be stored in the same KV storage device (e.g., 330 may be connected by a plurality of keynames located in common. Moreover, when batch writing is supported, object values can be batched together for better throughput.

In summary, the virtual device management layer 120, through the buffer, stores k recently stored keys (k) from k different KV storage devices 330, Objects 350 may be selected. The virtual device management layer 120 then returns the value object 450 (different from the largest value object (s) 470 in the parity group 440) corresponding to each of the key objects 350 And padded with virtual zero padding 350 to make the value objects 450 the same size (e.g., the size of the largest value object (s) 470). Virtual device management layer 120 may then generate r parity objects from k key objects 350 using an MDS code process. The virtual device management layer 120 then allocates r parity objects 460 to k KV storage devices 330 where the key objects 350 of the N KV storage devices 330 were selected And may write to the other r KV storage devices 330. At this time, k + r is equal to N. The virtual device management layer 120 may thereafter generate a metadata object representing the above-described information. Finally, the virtual device management layer 120 associates the key objects 350 and the parity objects 460 with the keys formed of the user key and the metadata identifier to the N KV storage devices 330 (e.g., Similar to a replication engine).

FIG. 5 is a block diagram of an embodiment of the present invention, in which a single object (k, r) erasure coding using typical erasure coding, or a reliability mechanism of "Splitting" Lt; / RTI > is a block diagram illustrating a group of KV storage devices configured to store key-value data in accordance with the present invention.

5, for values having larger value sizes than the value sizes of the values suitable for the reliability mechanisms of Packing using the above-mentioned Object Replication and typical erasure coding, Virtual device management layer 120 may select a reliability mechanism of single object (k, r) erasure coding, or "splitting" using typical erasure coding. The reliability mechanism of splitting using typical erasure coding has a relatively large value size and may be advantageous if the KV value 570 is divided into k equal-sized splits / chunks / values / objects 550 KV pair-of-trust mechanism that may be suitable for a KV value / object 570 that may have throughput.

After partitioning the KV value 570, the virtual device management layer 120 may calculate a checksum for each of the k objects 550, according to an embodiment. Thereafter, the virtual device management layer 120 may insert metadata before each of the k objects 550.

Splitting using typical erasure coding divides the KV value 570 into a plurality of smaller objects 550 followed by splitting a plurality of smaller objects 550 of the KV value 570 into k And spreading over successive storage devices 530. Thus, the size of k k-sized objects 550 may be supported by each of the underlying KV storage devices 530.

When using splitting using typical erasure coding, the virtual device management layer 120 may use systematic MDS code (e.g., the virtual device layer 120 as the base code, such as (4,2) Reed-Solomon code (K, r) MDS erasure coding), as well as the r parity values / objects 560 generated using the (k, r) MDS erasure coding. Thereafter, in a manner similar to the reliability mechanism of packing using the above-described typical erasure coding, the virtual device management layer 120 allocates k objects 550 and r parity objects 560 to N Lt; / RTI > storage devices (530). (k + r = N)

Therefore, the virtual device management layer 120 can divide a relatively large KV value 570 into k objects 550, calculate and add r parity objects 560, 550 and r parity objects 560 may be stored in k + r KV storage devices 530.

After hashing the key 580 corresponding to the KV value 570, the virtual device management layer 120 stores the corresponding object (e. G., ≪ RTI ID = 0.0 > (E.g., first of the k objects 550, in the embodiment of FIG. 5, D1 may be stored in the hash mark zero), the primary KV storage device 530a In the embodiment, KV-SSD2) can be determined. The k + r objects 550 and 560 may be written into the primary KV storage device 530a and the N-1 contiguous KV storage devices 530, respectively, under the same user key name. 5, the first of the k objects 550 can be written to the primary KV storage device 530a "KV-SSD2", and the rest of the k objects 550 KV-SSD3 "to" KV-SSDN "and" KV-SSD1 "along with the r parity objects 560 in the order of the cyclic manner. (E. G., In a manner analogous to that described for the reliability mechanism of object replication described above).

In summary, the virtual device management layer 120 may divide a relatively large KV object 570 into k equal sized objects 550, summarizing the reliability mechanism of splitting using typical erasure coding. The virtual device management layer 120 may then generate r parity objects 560 for k objects 550 using the MDC code process. The virtual device management layer 120 may then hash the key corresponding to the KV value 570 to determine the primary KV storage device 530a where the object will be located. The virtual device management layer 120 then creates a virtual KVM storage device 530a and N-1 contiguous KV storage devices 530, which are generated by the virtual device management layer 120 and in a circular fashion, It is possible to write the k + r objects 550 and 560 under the same user key name which may include an appropriate MetaID field.

Referring again to Figures 3 and 4, in accordance with another embodiment, the virtual device management layer 120 may use K-object (k, r, d) erasure coding or multi-object "Packing (E. G., According to flowchart 200 of FIG. 2). ≪ / RTI > The current reliability mechanism is similar to the reliability mechanism of packing using the previously described typical erasure coding in that the virtual device management layer 120 packs k objects into k KV storage devices. However, packing using regenerative erasure coding uses (k, r, d) regeneration codes instead of using typical (k, r) erasure codes. Thus, Figures 3 and 4 can be generally referenced for the current embodiment.

Thus, if the regeneration codes are suitable, but it is not feasible to partition the objects and it is more appropriate to preserve the objects, a packing using regenerative erasure coding may be used. Packing using regenerative erasure coding may be suitable for value sizes larger than the value sizes used for the previously mentioned object duplication and reliability mechanisms of the splitting and packing using typical erasure coding. Packing using regenerative erasure coding may be used when reading multiple subpackets of an object does not result in lower performance than reading the entire object. Packing using regenerative erasure coding may be used to prevent the underlying KV storage devices (e.g., KV storage devices 130 of FIG. 1, or KV storage devices 330 of FIG. 3) In the case of KV storage devices that are aware of the regeneration code.

Figure 6 is a block diagram of an embodiment of the present invention configured to store key-value data according to a reliability mechanism of single object (k, r, d) erasure coding, or "Splitting" using regenerative erasure coding Lt; / RTI > is a block diagram showing a group of KV storage devices.

6, the current reliability mechanism is similar to that shown in FIG. 4 except that the virtual device management layer uses (k, r, d) regeneration codes instead of the typical (k, r) MDS elimination coding Likewise, it can be made to operate in a manner similar to splitting using typical erasure coding. The current reliability mechanism, such as packing using regenerative erasure coding, may be suitable when KV storage devices 630 are KV storage devices that recognize the regeneration code that can assist during recovery / reconfiguration.

It may be desirable for object 670 to have a larger value size than the objects corresponding to the previously described reliability mechanisms and to read a plurality of subpackets 690 of k splits 680 of object 670, Splitting using regenerative erasure coding may be suitable if it does not result in lower performance than reading 680 (e.g., performed with the reliability mechanism of the splitting using typical erasure coding).

The reliability mechanism of splitting using regenerative erasure coding is that the objects are divided into k equal sized objects / splits 650 and splits 650 are split into multiple subpackets 690 (e.g., (Which is divided into four subpackets 690 per split 650) in the embodiment, and reading multiple subpackets 690 from the object 670 is virtually divided by the entire object 670 (KV pair) units that may be suitable for objects / KV values with very large value sizes that can have the appropriate throughput if they have better throughput than if they had better throughput than if they had read through. The value size is supported by all primary KV storage devices 630.

Similar to splitting using typical erasure coding, the virtual device management layer 120 of the current reliability mechanism, as shown in FIG. 4, can add r parity objects 660 using a systematic regeneration code And may write k splits 650 and r parity objects 660 to the N KV storage devices 630. (k + r = N). However, each of the r parity objects 660 may include a number of parity subpackets 692 (e.g., the number of subpackets 690 per split / k objects 605) Corresponding numbers). Unlike splitting using typical erasure coding, in the present embodiment the base code can be a (4,2,5) zigzag code.

By summarizing the reliability mechanism of splitting using regenerative erasure coding, the virtual device layer 120 can divide a large KV value 670 into k equal sized objects 650. The virtual device management layer 120 may then divide each of the k objects 650 into m equal-sized subpackets 690. m is an integer. The virtual device management layer 120 may then generate r parity objects 660 for k objects 650 using the regenerative coding process and each of the r parity objects 660 May be partitioned into m equal-sized parity subpackets 692. The virtual device management layer 120 may then hash the key corresponding to the KV value 670 to determine the primary KV storage device 630a where the object will be located. The virtual device management layer 120 then creates a virtual KVM storage device 530a and N-1 contiguous KV storage devices 530, which are generated by the virtual device management layer 120 and in a circular fashion, It is possible to write (k + r) objects 650, 660 including m subpackets 690, 692 for each under the same user key name that may include the appropriate MetaID field.

As described above, the virtual device management layer may select an appropriate reliability mechanism from the group of reliability mechanisms for storage of data based on one or more attributes of the data. Thus, the embodiments described herein provide an improvement in the field of memory storage because the described reliability mechanisms can each perform a single key recovery procedure. If all memory devices are failing, the virtual device management layer of an embodiment of the present invention can recover all of the keys present in the failed memory device and copy them to the new memory device. The virtual device management layer performs an iterative operation on all the keys existing in the memory devices adjacent to the failed memory device of the reliability group and performs a plurality of key unit operations on the keys that the reliability mechanism judges that the failed mechanism exists in the failed memory device , Recovery and copying of the entire keys can be achieved.

Very large KV pairs having larger value sizes than those supported by the underlying reliability mechanisms (e.g., in accordance with the underlying storage devices constraints) are explicitly partitioned into a plurality of KV pairs by the reliability manager, The embodiments described further provide an improvement in the field of memory storage, as they store a plurality of splits and split count information together with the metadata stored in the values.

The embodiments described herein, although specific terms are employed, are to be interpreted in a general and technical sense, without being limited thereto. In some instances, unless otherwise stated with respect to the illustrated embodiment, as indicated by ordinary skill in the art to which this invention pertains, features, attributes, and / or features described in connection with the specific embodiment The elements may be used in combination or independently with the described features, attributes, and / or elements in connection with other embodiments. It will therefore be appreciated by those skilled in the art that various changes in form or details may be made therein without departing from the spirit and scope of the following claims and the detailed description thereof as a functional equivalent thereof.

Claims (20)

A data storage method of a key-value reliability system including N (where N is an integer) storage devices grouped into a reliability group as a single logical unit and managed by a virtual device management layer,
Determining whether the data satisfies a threshold corresponding to a reliability mechanism for storing the data;
Selecting the trust mechanism if the threshold is satisfied; And
And storing the data in accordance with the selected reliability mechanism. The method according to claim 1,
Wherein the threshold is one or more of an object size of the data, a throughput consideration of the data, a read / write temperature of the data, and a basic erasure coding capability of the N storage devices Based method. The method according to claim 1,
Further comprising testing the data for the trust mechanism using one or more Bloom filters or caches. The method according to claim 1,
The selected reliability mechanism, one or more checksums for each of the N storage devices storing the data, object sizes of values of the data stored in each of the N storage devices storing the data, Further comprising inserting metadata for recording the location of parity group members of the N storage devices together with a key corresponding to the data indicating to which of the N storage devices the data is stored . The method according to claim 1,
Wherein the selected trust mechanism comprises object replication,
Wherein the step of storing the data comprises:
Selecting a KV value;
Computing a hash for hashing the key corresponding to the selected KV value;
Determining some of the N storage devices to store replicas of key objects corresponding to the KV value; And
And writing the updated values corresponding to the KV value to each of the determined partial storage devices under the same user key name. The method according to claim 1,
Wherein the selected reliability mechanism comprises packing,
Wherein the step of storing the data comprises:
Selecting k key objects stored in k (where k is an integer) storage devices among the N storage devices of the reliability group;
Retrieving k value objects corresponding to the k key objects;
Padding virtual zeroes on the ends of the value objects having the largest value size among the k value objects so that the virtual sizes of all of the k value objects become equal;
Generating r (where r is an integer) parity objects from the k key objects;
Writing the k key objects into the k storage devices; And
And writing the r parity objects to r storage devices among the N storage devices,
Wherein each of the r storage devices is distinguished from the k storage devices, with the proviso that k + r = N. The method according to claim 6,
The selected reliability mechanism includes packing using typical erasure coding,
Wherein the N storage devices are configured with typical (k, r) maximum distance separable (MDS) erasure coding. The method according to claim 6,
Wherein the selected reliability mechanism comprises packing using regenerative erasure coding,
Wherein the N storage devices comprise (k, r, d) regenerative erasure coding. The method according to claim 1,
Wherein the selected reliability mechanism comprises a splitting,
Wherein the step of storing the data comprises:
Selecting a KV value;
Dividing the KV value into k (where k is an integer) objects of the same size;
Generating r (r, integer) parity objects from the k equal-sized objects;
Computing a hash for hashing the key corresponding to the selected KV value;
Determining a primary device on which the KV value is to be located among the N storage devices based on the hash; And
Writing the k number of identical sized objects and each of the r parity objects to the N storage devices in a sequential order starting from the main device, with the proviso that when k + r = N . 10. The method of claim 9,
The selected reliability mechanism includes splitting using typical erasure coding,
Wherein the N storage devices are configured with typical (k, r) maximum distance separable (MDS) erasure coding. 10. The method of claim 9,
Wherein the selected reliability mechanism comprises packing using regenerative erasure coding,
Wherein the N storage devices comprise (k, r, d) regenerative erasure coding,
Wherein the step of storing the data comprises:
Dividing the k equal sized objects into m (where m is an integer) subpackets using the regenerative erasure coding; And
Further comprising partitioning each of the r parity objects into m parity subpackets. A data reliability system for storing data based on a selected reliability mechanism,
N (where N is an integer) storage devices configured as virtual devices using stateless data protection; And
And a virtual device management layer configured to manage the N storage devices as the virtual device and store data in selected ones of the N storage devices according to the selected reliability mechanism,
The virtual device management layer comprising:
Determining if the data satisfies a threshold corresponding to a reliability mechanism for storing the data;
Select the trust mechanism if the threshold is satisfied;
And store the data in accordance with the selected trust mechanism. 13. The method of claim 12,
Wherein the selected trust mechanism comprises object replication,
The virtual device management layer comprising:
KV value is selected;
Calculate a hash for hashing the key corresponding to the selected KV value;
Determining some storage devices of the N storage devices to store replicas of key objects corresponding to the KV value;
And to write the updated values corresponding to the KV value to each of the determined some storage devices under the same user key name. 13. The method of claim 12,
Wherein the selected reliability mechanism comprises a packing,
The virtual device management layer comprising:
Selecting k key objects stored in k (where k is an integer) storage devices among the N storage devices;
Retrieving k value objects corresponding to the k key objects;
Padding virtual zeroes at the ends of the value objects having the largest value size among the k value objects such that the virtual value sizes of all of the k value objects become equal;
Generating r (where r is an integer) parity objects from the k key objects;
Write the k key objects into the k storage devices;
And storing the data by writing the r parity objects to r storage devices of the N storage devices,
Wherein each of the r storage devices is distinct from the k storage devices, and k + r = N. 13. The method of claim 12,
Wherein the selected reliability mechanism comprises a splitting,
The virtual device management layer comprising:
KV value is selected,
Dividing the KV value into k (where k is an integer) objects of the same size;
Generating r parity objects from k equal sized objects, where r is an integer;
Calculate a hash for hashing the key corresponding to the selected KV value;
Determining a main device on which the KV value is to be located among the N storage devices based on the hash;
Write the k pieces of the same size objects and the r parity objects to the N storage devices in a sequential order starting from the main device,
Where k + r = N. 16. The method of claim 15,
Wherein the selected reliability mechanism comprises splitting using regenerative erasure coding,
Wherein the N storage devices comprise (k, r, d) regenerative erasure coding,
Wherein the virtual device management layer divides the k pieces of the same size objects into m (where m is an integer) subpackets using the regenerative erasure coding, and allocates each of the r parity objects to m parity subpackets The data storage system further configured to store the data. And computer code in which a method of storing data in a key-value reliability system, which is implemented as a single logical unit, is grouped into a reliability group and includes N storage devices managed by a virtual device management layer, For non-transient computer readable media,
The method comprising:
Determining whether the data satisfies a threshold corresponding to a reliability mechanism for storing the data;
Selecting the trust mechanism if the threshold is satisfied; And
And storing the data in accordance with the selected reliability mechanism. 18. The method of claim 17,
Wherein the selected trust mechanism comprises object replication,
Wherein the step of storing the data comprises:
Selecting a KV value;
Computing a hash for hashing the key corresponding to the selected KV value;
Selecting some of the N storage devices to store a replica of key objects corresponding to the KV value; And
Writing the updated values corresponding to the KV value to each of the determined some storage devices under the same user key name. 18. The method of claim 17,
Wherein the selected reliability mechanism comprises a packing,
Wherein the step of storing the data comprises:
Selecting k key objects stored in k (where k is an integer) storage devices among the N storage devices of the reliability group;
Retrieving k value objects corresponding to the k key objects;
Padding virtual zeroes at the ends of the value objects having the largest value size among the k value objects such that the virtual value sizes of all of the k value objects become equal;
Generating r (where r is an integer) parity objects from the k key objects;
Writing the k key objects to the k storage devices; And
And writing the r parity objects to r storage devices among the N storage devices,
Wherein each of the r storage devices is distinct from the k storage devices, and k + r = N. 18. The method of claim 17,
Wherein the selected reliability mechanism comprises a splitting,
Wherein the step of storing the data comprises:
Selecting a KV value;
Dividing the KV value into k (where k is an integer) objects of the same size;
Generating r (where r is an integer) parity objects from the k equal-sized objects;
Computing a hash for hashing the key corresponding to the selected KV value;
Selecting a primary device on which the KV value is to be located among the N storage devices based on the hash; And
Writing the k number of identical sized objects and each of the r parity objects to the N storage devices in a sequential order starting from the main device,
k + r = N. < / RTI >

Images