DE102013114069A1 - Memory system for changing operating characteristics of storage device i.e. solid state drive, has storage including adaptation controller to receive command from configuration controller and to determine whether to enable feature - Google Patents

Abstract

The system has a host (100) including a configuration controller (110) to receive an input command and to output a configuration command corresponding to the input command. A storage (200) includes multiple features and an adaptation controller (210) to receive the configuration command from the configuration controller and to determine whether to enable each of the features, where the first feature drives the storage to preferentially process a read command, and the second feature drives the storage to preferentially process a write command among commands input to the storage. An independent claim is also included for a storage of a memory system.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a continuation-in-part application of patent application Serial No. 13733309 entitled "Memory system using a memory having firmware with a plurality of functions" filed on January 3, 2013, which is incorporated herein by reference USC § 119 the priority of Korean Patent Application No. 10-2012-0005764 filed with the Korean Intellectual Property Office on January 18, 2012, and all the benefits that result under 35 USC § 119, the contents of both applications being incorporated herein by reference.

BACKGROUND

Modern memory systems typically include a plurality of standalone memory devices that are passive and whose performance characteristics are generally fixed when manufacturing is complete. A processor (eg, RAID controller) that executes software (firmware) is necessary to add intelligence to cause the accumulation of unintelligent memory devices to operate as a unit. Since storage devices such as Solid State Drives (SSDs) can also be controlled by the software or firmware, efforts have been made to control the operating characteristics of a solid state drive according to the environment of use.

Traditional approaches to completing user qualification tests require a manufacturer to provide last-minute engineering operations to customize the storage devices for each buyer. Buyers typically wish to recalibrate their software systems whenever new models of storage devices are used because the characteristics of the storage devices are largely heterogeneous. However, the requirements for a storage device are often not valid with another device. As a result, conventional adaptation approaches can not be maintained in part because the manufacturer's development cost increases with the number of storage devices and the purchasers in need of adjustment.

Thus, a framework that enables easy reconfiguration of storage systems on behalf of buyers is important. For example, solid state drive (SSD) optimization software, such as Samsung's Magician ^™, adjusts the performance of SSDs for a buyer's system. However, buyers have very limited optimization options, the optimization metric is device-oriented rather than user-oriented, and optimization is uncontrolled or quantifiable. Additionally, memory device characteristics may change over time due to degradation of the storage media, such as wear and aging. This may violate the initial assumption that the buyer had, which can not easily be detected, until user-level malfunctions occur.

Another type of reconfigurable storage device operation allows a buyer to select individual functions to configure a storage device. In this approach, instead of adapting a buyer's system to a new storage device, a reconfigurable storage device allows the buyer to customize the storage devices to their systems, which simplifies maintenance and an upgrade process.

Although reconfigurable memory devices can provide more flexibility in performance optimization and allow buyers to make an adjustment, several challenges remain. One challenge is that the recalibration process is a combinatorial problem whose complexity increases exponentially with the number of functions of the memory device to be adapted. In other words, current approaches do not provide a systematic configuration method for function selection. For example, when a buyer changes the value of three functions, it may be difficult for the buyer to determine what effect the combination of functions will have on the performance of the storage device.

A similar challenge is that the selection of functions by the buyer is accomplished by a software user interface in which the buyer manually selects the functions. Manual selection of functions without a systematic configuration procedure or performance guideline results is essentially a trial and error process.

Finally, the conventional reconfiguration process does not address the effects of memory device characteristics which change over time due to the degradation of the storage medium. Such changes may not continue to validate the original selection of features for a particular consumer environment.

As a result, the software-defined storage (SDS) trend in which storage resources required by an application can be software-defined and automatically retained requires an improved reconfigurable storage process that is more flexible.

SHORT VERSION

The exemplary embodiment provides methods and systems for reconfiguring storage devices. Aspects of an example environment include receiving user request information for a storage device and automatically generating function settings for the storage device from the user request information and a device profile for the storage device; and using the function settings to automatically reconfigure the storage device into one or more logical devices that have independent performance characteristics. Further exemplary embodiments include receiving feedback from the memory device regarding runtime performance data of at least one of the memory device and the logical devices; and in response to a determination that the runtime performance data does not satisfy the user request information, dynamically adjusting the configuration of the one or more logical devices to satisfy the user request information.

In accordance with the method and system disclosed herein, the exemplary embodiments provide a reconfiguration process that eliminates the need to modify configuration software on a host to configure different storage devices per user request. The exemplary embodiments are in contrast to conventional methods in which the host's configuration software would need to be modified to accommodate different memory devices. Additionally, as the exemplary embodiments automatically translate the user requirements into functional settings, the need for the user to manually select combinations of functions when configuring the memory device is eliminated.

BRIEF DESCRIPTION OF SOME VIEWS OF THE DRAWINGS

These and / or other features and advantages of the present general inventive concept will become more readily apparent from the description of the embodiments which follows, taken in conjunction with the accompanying drawings, in which:

1 Fig. 12 is a block diagram illustrating an exemplary embodiment of a reconfigurable memory system;

2A and 2 B Are flowcharts illustrating an embodiment of a process for reconfiguring a storage system;

3 Fig. 10 is a flowchart illustrating in further detail the process for configuring the reconfigurable storage device;

4 Fig. 10 is a flowchart illustrating an exemplary conversion process performed by the configuration planner to generate a plan;

5 Figure 3 is a flowchart illustrating processing of the configuration component or configuration component on the reconfigurable storage device;

6 FIG. 10 is a flowchart illustrating an execution flow of the configuration advisor according to an embodiment; FIG. and

7 Figure 12 is a block diagram illustrating another embodiment for a reconfigurable memory system.

DETAILED DESCRIPTION

Reference will now be made in detail to the embodiments of the present general inventive concept, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to like elements throughout. The embodiments are described below to explain the present general inventive concept while referring to the figures.

Advantages and features of the present invention and methods for achieving the same may be more readily understood by reference to the following detailed description of embodiments and the accompanying drawings. However, the present inventive concept may be embodied in many different forms and should not be considered as limited to the embodiments discussed herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and the concept of the general inventive concept will be fully conveyed to those skilled in the art, and the present generic inventive concept will be defined only by the appended claims. In the drawings, the thicknesses of layers and regions are exaggerated for clarity.

The use of the terms "one" and "one" and "the" and similar references in the context of describing the invention (particularly, in the context of the following claims) must be considered to include both the singular and the plural, as long as so not otherwise indicated herein or clearly contradicted by the context. The terms "having," "having," "including," and "containing" must be considered open-ended (i.e., meaning "including but not limited to") unless otherwise noted.

As used herein, the term "component" or "component" or "module" means, but is not limited to, a software or hardware component such as a field programmable gate array (FPGA) = Field Programmable Gate Array) or an application specific integrated circuit (ASIC), which performs certain tasks. A component or module may be advantageously configured to reside in the addressable storage medium and configured to execute one or more processors. Thus, a component or module may illustratively include components such as software components, object-oriented software components, class components and task components, operations, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases , Data structures, tables, arrays, and variables. The functionality provided for the components and components or modules may be combined into fewer components or components or modules, or further separated into additional components and components or modules.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It should be understood that the use of any and all examples or exemplary terms provided herein is intended merely to better illuminate the invention, and not to limit the scope of the invention, unless otherwise specified. Furthermore, unless otherwise defined, all terms defined in commonly used dictionaries should not be overly interpreted.

1 FIG. 10 is a block diagram illustrating an exemplary embodiment of a reconfigurable memory system. FIG. The system has a host 100 on which with a reconfigurable storage device 102 is coupled. In one embodiment, basic components of the host 100 at least one processor 104 , a store 106 and a host controller or host controller 108 exhibit. In an embodiment, basic components of the reconfigurable memory device 102 a memory controller or memory controller 111 , a main memory 112 , at least one processor 114 and a memory 116 exhibit.

The main memory 112 implements native or built-in storage functionalities and sees the storage space for the reconfigurable device 102 in front. The main memory 112 can one nonvolatile memory such as a flash (implemented, for example, as a solid state drive (SSD)), a PCRAM, an RRAM, an STT RAM, and the like, and a memory management firmware (not shown) such as FTL (Flash Translation Layer). , GC (Garbage Collector) and the like, but is not limited thereto. One or more logical devices 121 can go to a section of main memory 112 contribute. The actual physical space required for the logical devices 121 depends on space or space efficiency requirements. A thin-provisioned logical device may occupy less physical memory space than a thick-provisioned logical device.

During normal operation, the host controller may 108 commands 124 which are associated with a content read / write operation to a memory controller 111 spend to the host's 100 to allow content to be stored in a main memory 112 are stored, processed, or new content in main memory 112 save. Here can be processing the content in the host 100 displaying or printing the contents in the form of images and documents, reproducing the contents in the form of audio and / or video and installing or executing the contents in the form of an application such as a computer program on a functional entity such as the processor 104 exhibit.

The memory controller 111 receives the command associated with the contents from the host controller 108 and controls content 126 which are in main memory 112 stored in accordance with the received command. For example, if the received command is a read command, the memory controller 111 the content 126 which are in main memory 112 are stored for the host 100 provide, and if the received command is a write command, the memory controller 111 new content 126 from the host 100 receive and the received contents in the main memory 112 to save.

Furthermore, although the host controller 108 and the memory controller 111 shown as separate components, in an alternative embodiment, the host controller 108 and the memory controller 111 be implemented as an integrated unit, if necessary.

As described above, a problem with the conventional memory devices is that they are fixed devices with little built-in intelligence. As a result, the process of adapting the storage device to a buyer request may be costly and time consuming.

According to the exemplary embodiment, the host is 100 continue with a configuration translation component 110 provided while the reconfigurable storage device 102 with a configuration component 118 , a feature set 120 and a device profile 122 is provided. In one embodiment, the configuration translation component 110 on the host 100 be configured so that they user requirements 123 which behavioral characteristics for the reconfigurable storage device 102 describe, receive and in a configuration command 128 which function settings for the behavioral characteristics for the reconfigurable storage device 102 Are defined. In one embodiment, functions of the configuration translation component 110 by separate components such as a configuration planner 136 , a device programmer 136 and a configuration advisor 138 be performed.

Consequently, the configuration component 118 on the reconfigurable storage device 102 be configured, so that they have the configuration command 128 receives and the function settings according to the configuration command 128 implementing, thereby, behavioral characteristics of the reconfigurable storage device 102 automatically configured. Thereafter, the memory controller 110 content 126 to / from the host 100 send / receive while the main memory 112 is driven or operated according to the set behavioral characteristics.

In another aspect of the exemplary embodiment, during operation of the reconfigurable memory device 102 the configuration translation component 110 a feedback or a feedback 130 from the reconfigurable storage device 102 concerning performance characteristics of the reconfigurable storage device 102 receive and can the configuration command 128 (and consequently the function set) accordingly based on the feedback 130 to adjust.

In one embodiment, the configuration translation component is 110 and the configuration component 118 implemented as software components. In another embodiment, the Components should be implemented as a combination of hardware and software. Although the configuration translation component 110 and the configuration component 118 As shown as individual components, the functionality of each can be combined in a smaller or larger number of modules / components. For example, in one embodiment, the memory controller 111 and the host controller 108 both on the host 100 be implemented as an integrated component or as separate components.

Likewise, although the user request information 123 as the host 100 supplied, in an alternative embodiment, the reconfigurable memory device 102 be adapted to the user request information 123 to receive and the user request information 123 for the host 100 provided.

The reconfigurable storage system of the exemplary embodiments may be applied to a wide range of storage markets from a customer to a business which stores on a single standalone machine disk (such as desktop, laptop, workstation, server, and the like) a storage device or a storage array, software-defined storage (SDS), application-specific storage, a virtual machine (VM), a virtual desktop infrastructure (VDI), a content Distribution network (CDN = Content Distribution Network) and the like.

In an embodiment, the reconfigurable memory device 102 For example, integrated into a semiconductor device may be a Personal Computer Memory Card International Association (PCMCIA), a Compact Flash (CF), a Smart Media Card (SM, SMC), a Memory Stick, a Multimedia Card (MMC, RS -MMC, MMCmicro), an SD card (SD, miniSD, microSD, SDHC), a universal flash memory (UFS = Universal Flash Storage) device and the like.

For example, in one embodiment, the main memory part 112 be formed of a plurality of nonvolatile memory chips, that is, a plurality of flash memories. As another example, the main memory part 112 of nonvolatile memory chips of different types (e.g., PRAM, FRAM, MRAM, etc.) instead of flash memory chips. Alternatively, the main memory part 112 may be formed of volatile memory, ie, DRAM or SRAM, and may have a hybrid type where two or more types of memories are mixed.

The 2A and 2 B FIG. 10 are flowcharts illustrating one embodiment of a process for reconfiguring a storage system. FIG. 2A FIG. 10 is a flow chart illustrating a configuration process for reconfiguring the reconfigurable storage device as required by a user 2 B Fig. 10 is a flow chart illustrating the process for configuring the reconfigurable storage device 102 during operation of the reconfigurable storage system.

Referring to 2A The process may begin by having the configuration translation component 110 Users request information 123 for a storage device and automatically receives function settings for the storage device from the user request information and a device profile for the storage device (block 200 ) generated. The configuration component 118 can then use the feature settings to automatically place the storage device in one or more logical devices 121 which have independent behavioral characteristics to reconfigure (Block 202 ).

Referring to 2 B FIG. 10 is a flow chart illustrating the process for configuring the reconfigurable storage device. FIG 102 during operation of the reconfigurable memory system. In this embodiment, the configuration translation component receives 110 feedback 130 from the storage device regarding runtime performance data 134 at least one of the storage device and the logical devices (block 204 ). In response to a determination that the runtime performance data does not satisfy the user request information, the configuration translation component fits 110 dynamically configure the one or more logical devices to satisfy the user request information (block 206 ).

An advantage of the reconfigurable storage system is that the memory management of the configuration translation component 110 (ie the user software) can be simplified via newer configurable devices, as the example embodiments have peculiarities and the Abstract reconfiguration complexity of storage devices. Another advantage is that the user software using new configurable storage devices can remain virtually constant regardless of models of storage devices because the user software can dynamically define and generate its own logical device from the reconfigurable storage devices according to the user requirements ,

Referring again to 1 According to another aspect of the exemplary embodiment, the user request information 123 a high-level description or description at high level of the behavioral characteristics of the reconfigurable memory device 102 as opposed to current values for different functions. According to one embodiment, the high-level description may be a service level agreement (SLA) for the reconfigurable storage device 102 exhibit. SLAs are commonly service contracts or service agreements concluded between two parties, such as a buyer and a manufacturer.

According to another embodiment, the high-level description may further include quality of service (QoS) requirements, which in one embodiment are part of the SLA. In another embodiment, however, the QoS may be separate from the SLA if the user is only interested in a performance aspect. QoSs, in the exemplary embodiments, may include performance characteristics of the memory device in terms of latency, IOPS, bandwidth, jitter, WAF / RAF, which are for defining the detailed performance expectancy of the user. For example, after a virtual device is created, it may be desirable to adjust only the QoS rather than creating a new logical device. An automatic configuration of the reconfigurable storage device 102 using the SLA and QoS as disclosed herein results in the configuration being accomplished in a quantitative manner.

In one embodiment, an SLA may have specifications for any combination of attributes expressed in an XML format. An attribute may be performance characteristics (QoS) of a storage device such as latency and IOPS or non-performance characteristics such as capacity, addressing mode and protection. Commitment is the average percentage of QoS compliance of operations. The reaction is what the host expects in the event of a QoS violation. The capacity specification may specify an amount of data space and space savings. The capacity may indicate a lot of data in the user space, not the actual storage space allocated in the device. In this sense, the capacity specification is independent of the mechanism that the device is currently using to implement it, but all that is needed is a guarantee that the amount of data is stored at the user level can be. For example, an apparatus employing a thick provisioning technique allocates the current space when a request is granted, while a device employing a thin provisioning technique allocates the space can not allocate in advance. Similarly, the actual physical space a device allocates with a space-saving function, such as over-commitment, compression, de-duplication, may be much smaller than the data space that the user requests.

The addressing mode specification may specify the manner in which data is to be addressed in memory: byte-addressable, word-addressable, and / or block-addressable. For example, a Peripheral Component Interconnect Express (PCIe) device is byte-addressable as well as block-addressable. Similarly, a single address space operating system can emulate byte addressing via a block addressable device. The QoS should be interpreted in the context of a particular addressing mode.

The protection specification may specify a security and access control. A device may support encryption and the user may use a particular encryption algorithm (e.g., 128-bit RSA, 256 AES) for its security requirements. The user may also require secure deletion of data with the risk of irreversibility to the user if the data is deleted. In addition, the user can define access controls such as read-only, write-once, write-once, at the query level. For example, a recovery image of the system may be stored in a write once logical device.

The performance specification may specify latency, data throughput, bandwidth, and / or input / output operations per second (IOPS) measurements, and so the user may have detailed performance requirements for a device. For example, the storage devices used to store video clips in a video-on-demand server or video-on-demand server must have high sequential read performance. In contrast, a storage device used to manage metadata may require high IOPS (and low latency).

Exemplary SLAs may include the following: Example 1) <sla>

In the above example, the user may be interested in a 100 GB device that is byte addressable. In addition, the 4 KB page read latency must be less than 10 ms, and the 4 KB write latency must also be less than 10 ms. Finally, the IOPS of the device must be at least 1K.

Example 2) <sla>

Example 3) <qos>

The reconfigurable storage device 102 may further comprise a plurality of operating functions, to which here as a function set 120 Reference is made. In one embodiment, the functional set 120 a hardware operation (eg, a clock frequency, an excessive reset ratio), an instruction termination strategy (eg, read before write, random access before sequential access, small request before large request, a scheduling algorithm, read before GC), logical device characteristics ( for example, capacity, queue length, time-out control, protection), a mapping strategy of logical addresses on physical addresses (for example, page mapping, block mapping, hybrid mapping), a background job execution strategy ( for example, garbage collection algorithm, period and frequency of garbage collection, background TRIM, background compaction, etc.), which are not exhaustive.

If the configuration translation component 110 the user request information 123 can receive the configuration translation component 110 automatically the user request information 123 translate into function settings which can be used to set specific functions in the function set 120 to activate / deactivate, as well as to specify values for certain functions. The effect of a function can be global for all logical devices 120 or local to a particular logical device. For example, adjusting the operating frequency of the device is global, which is the behavior of all logical devices 121 changes while increasing or increasing the queue length of a logical device is only local to the device.

According to one aspect of the exemplary embodiment, the configuration translation component generates 110 the function settings based on both the user request information 123 as well as the device profile 122 ,

In one embodiment, the device profile 122 Performance ranges 131 of functions (for example, the range of an adaptable clock frequency), in particular for the memory device, performance models 132 and runtime performance data 134 exhibit. The runtime performance data 134 may have transit time performance statistics measured by the device during operation.

According to an example embodiment, the performance models 132 and functions of the reconfigurable memory unit 102 which have a correlation with the device performance. In one embodiment, the performance models 132 have mathematical models (eg, linear, polynomial, logarithmic, exponential or step function) for individual functions and / or combinations of function settings and / or ideal performance data measured in a controlled environment to estimate the QoS.

Table 1 below shows an exemplary set of functions 120 and associated performance models. The models in this invention are exemplary and depend on suppliers and devices. function effect Performance model clock speed IOPS Model W (for example, linear) queue size IOPS Model X (for example Log) thermal IOPS / latency Model Y (for example, inverse polynomial) Over-Provisioning latency Model Z (eg Log) time-out N / A Constant TABLE 1

In one embodiment, any function that may change the performance characteristics of the device may be an associated performance model 132 and / or have ideal performance reference tables if a model is impractical. In one embodiment, either the configuration translation component 110 or the configuration component 118 the device profile 122 and / or the performance models 132 to determine if a specification of the user request information 123 through the reconfigurable storage device 102 can be fulfilled. For example, a user may improve the response time of a read operation by increasing the clock rate of the physical device. However, a device has a certain range of clock rates that can be adjusted. When the device is already reaching its maximum clock rate, the clock rate adjustment function can not be used to improve latency. Similarly, if IOPS can be modeled at clock rate as a linear function to a certain extent, for example, how much the clock rate is to be increased can be determined.

In one embodiment, the device profile 122 which the performance models 132 comprising the configuration translation component 110 be made available, for example by the reconfigurable storage device 102 the configuration translation component 110 be reported. In another embodiment, the device profile 122 on the host 100 or in a remote place, on which by the host 100 can be accessed, and / or the reconfigurable storage device 102 be saved. Although the performance models 132 as part of the device profile 122 stored in an alternative embodiment, the performance models 132 separate from the device profile 122 be saved.

If a provider of the reconfigurable storage device 102 the device profile 122 to the host 100 can explain the configuration translation component 110 in the host 100 the performance areas 131 , the performance models 132 and the runtime performance data 134 use effectively. Otherwise, the configuration translation component 110 Access to their own device profile 122 for each type of reconfigurable storage device 102 need.

In one embodiment, the runtime performance data may be 134 , the requirement profile 122 typical performance indices that are available through a standard interface such as SMART (Self-Monitoring, Analysis and Reporting Technology) in the ATA specification, and vendor-specific runtime performance metrics such as min / max / avg resp Minimum / maximum / average read / write latency, minimum / maximum / average queue length and wait time, minimum / maximum / average internal IOPS, a read / write gain ratio, garbage collection Time or garbage collection time, a timeout count and the like.

In one embodiment, other information may be provided as part of the runtime performance data 134 are stored, but not limited to: device status, device background job status, and host background job status.

The device status may include status information about hardware and software of the reconfigurable storage device 102 exhibit. This may include not only the global information such as the degree of deterioration of the storage medium (wear level), a supercapacitor state, a device temperature, an over-provisioning ratio (or amount), and the like, but also the local ones Information of the logical device such as a device size, a device capacity, a device usage, an SLA, a function set, function settings, the degree of data fragmentation, a timeout, a protection, an addressing mode, etc.

The device background job status may include a status of device background jobs determined by device policies, but not visible to outside the device, such as background TRIM, background garbage collection, background compaction, and others.

The host background job status may include the status of the device background jobs triggered by the host, such as the offline test of the SMART self-test suite in the ATA specification.

The configuration translation component 110 in the host 100 , what a 1 1, the high-level user QoS requirements may be converted to SLA device-specific configuration commands to configure the reconfigurable memory device 102 to change. The configuration translation component 110 can the relationship between the user requirements 123 and operations / strategies of the reconfigurable storage device 102 , which are device-specific, understand. For example, a storage device may require a queue with a large queue depth to meet a particular IOPS request. Similarly, low-end devices may require a queue with a large queue depth and a read priority to achieve a short read latency, while high-end devices may achieve the same latency without a read priority.

3 FIG. 10 is a flowchart which, in further detail, illustrates the process of configuring the reconfigurable storage device. FIG 102 illustrated. In one embodiment, the process steps performed on the host 100 be performed by the configuration translation component 110 and the process steps performed on the reconfigurable storage device 102 are performed by the configuration component 118 carried out.

The process may be on the host 100 begin by receiving an SLA and receiving the device profile 122 (Block 300 ). For each SLA QoS request, configurable feature settings are based on the device profile 122 the reconfigurable storage device 102 determined (block 302 ) and possible combinations of function settings (QoS) are determined for the SLA (block 304 ). The combination of feature settings will be based on the device profile 122 for the reconfigurable storage device 102 prioritized (block 306 ). For each of the combination of function settings, a set of configuration commands is referred to as a plan is generated (block 308 ). The plan or individual configuration commands of the plan become a reconfigurable storage device with a transaction ID 102 transferred (block 310 ). A plan may be rejected by the configuration component in the device because the device state may change in the case of races while creating tarpaulins. Accordingly, the host can 100 Plan iteratively until one or none of the plans succeed (Block 312 ).

The process may be on the storage device 102 by receiving the plan from the host (block 320 ). The plan is made by checking the availability and applicability of each feature based on the device profile 122 validated (block 322 ). All functions in the plan are automatically set, and in the event of a malfunction, any changes are undone (block 324 ). One or more logical devices are generated based on the function settings (block 326 ). The device profile 122 we updated per configuration of the plan (block 328 ). A completion response then becomes the host 100 sent (block 330 ).

During operation of the storage device 102 the plan is compared to the device profile, in particular the real-time performance statistics 134 (Block 332 ). A notification will then be sent to the host as feedback 130 sent if the device profile does not meet the plan (block 334 ). The configuration advisor 138 then uses the feedback to guide the configuration planner to a proper fit ( 6 ).

The following describes in more detail the function of the three components that comprise the configuration translation component 110 as in 1 have shown - the configuration planner 136 , the configuration advisor 138 and the device programmer 136 as well as the configuration component 118 in the reconfigurable storage device 102 ,

The configuration planner 136 receives the user requests 123 (for example, the SLA) and analyzes possible configuration plans, which configuration commands 128 based on the strategies that the device supports, as in the device profile 122 (for example, the runtime performance data 134 ) and generates a configuration plan as a recommendation. More specifically, the configuration planner 136 a conversion algorithm to convert SLA specifications to function settings based on the device profile 122 It determines optimal combinations from the converted function settings and generates a plan which configuration commands 128 has, from the optimal combinations.

As an example, assume that the storage device has a device profile as follows: latency increases by x% and IOPS decreases by y% with an encryption function, while prioritizing read before write improves read latency by z%, and the increase in queue depth IOPS improved by w% ratio. If the user wants to achieve the same performance even with an encryption option, the configuration scheduler can 136 must enable a read priority and / or increase the queue depth to compensate for the latency increase due to encryption.

The configuration planner 136 can generate multiple configuration plans from the user requests, and the configuration planner 136 chooses the best plan, specifying the performance and status of the storage device 102 and the user requirements. The configuration planner 136 can have multiple performance models 132 which are associated with each function of the reconfigurable memory device, have a correlation and anti-correlation between functions and other and multiple scheduling algorithms representing the performance models 132 evaluate and generate the configuration plans. The configuration planner 136 can the performance models 132 with the storage device 102 use together when the storage device 102 Access to the performance profiles 132 provides. Otherwise, the host may 100 have their own performance profiles.

If the configuration component 118 in the storage device 102 the plan receives the configuration component 118 reject the plan if the plan exceeds the capability that the device can support. For example, if the device does not provide performance profiles, or if the device experiences any delays or sequences when capturing the current performance characteristics, the performance profile provided by the device may be stale. In this case, the device can not reward the plan, even if the decision on their performance profile 122 is based. In one embodiment however, the configuration component 118 enforce or enforce the plan at your own risk as well.

The configuration plan may include a set of configuration commands, each of which is a predetermined vendor-unique command or a standard command to which the storage device 102 can understand. For the previous example, the configuration planner must be 136 To send two commands: one for the scheduling strategy, which gives a priority to read before write, and another for the queue properties, which increases the queue size. Optionally, the configuration planner 136 send a command for the entire plan with a transaction ID. Depending on the settings and status of the storage device 102 can the configuration component 118 the configuration commands or the plan reward or not.

4 FIG. 3 is a flowchart illustrating an exemplary conversion process performed by the configuration scheduler 136 is performed to generate a plan. The process corresponds to the blocks 300 to 308 of the 3 , When the SLA is received (block 400 ) determines if multiple QoS requests are to be resolved (block 402 ). If so, possible combinations of feature settings that can satisfy a current QoS request are identified (Block 404 ). The decision in this step is more or less a general policy based on the static information (previously defined knowledge), which may vary depending on devices, but may reduce the search space of the next steps. It is then determined whether a combination of function settings satisfies the current QoS request with a particular device of interest based on the performance profile (Block 406 ). If not, the next combination of function settings is obtained (block 408 ). It is then determined if the combination is empty to confirm that all possible combinations are attempted in the event of malfunction (block 410 ), and if so, an error is generated (block 412 ). If the combination is not empty, the process moves to block 406 continued.

If the combination of function settings meets the current QoS requirement (block 406 ), then it is determined if the combination of function settings is combined with the already selected configurations for all the SLAs and QoSs of that device (Block 414 ). Again, due to the dynamic change in device performance in the case of races, the combination may not work. If so, a function settings configuration command is added to the plan (Block 416 ) and the next unresolved QoS in the SLA is obtained (block 418 ). If it is determined that there are no longer QoS requirements for the operation (block 402 ), gives the configuration planner 136 the plan to the configuration component 118 in the reconfigurable storage device 102 off (block 420 ).

Referring again to 1 the device programmer receives 140 the plan, which by the configuration planner 136 is selected and sends the plan to the reconfigurable storage device 102 via a configuration command channel (configuration command channels) for executing the plan and configuration of the memory device accordingly. The device programmer 140 may receive a response from the reconfigurable storage device regarding execution of the plan.

Because the configuration planner 136 several plans recommend (which are probably arranged by the layout quality), the device programmer can 140 the plans in case of a negative response from the reconfigurable storage device 102 until a plan is found which is actually with the reconfigurable storage device 102 is working. If the device programmer 140 any errors during schedule programming (for example due to the rejection of the configuration component) 118 ), the device programmer should 140 be able to withdraw the changes made by the plan, if necessary.

The configuration command channel (s) used to execute the configuration commands and the feedback channel (feedback channels) which are used to monitor and collect the performance profile and status information of the device , may be a standard host interface such as SATA, SAS, PCIe and NVMe or a special interface designed for this purpose.

5 FIG. 10 is a flowchart showing a processing of the configuration component. FIG 118 on the reconfigurable storage device 102 illustrated. This process corresponds to block 202 of the 2A . where the reconfigurable storage device using the function settings from the host 100 is reconfigured, as in the blocks 320 to 328 of the 3 ,

The process can begin when the configuration component 118 a plan from the device programmer 140 on the host 100 receives which at least one configuration command 128 which contains function settings (block 500 ).

The configuration component 118 can first check the availability and applicability of the feature settings in the plan based on the device profile (block 502 ). Checking for availability may include determining whether the storage device supports the feature settings by comparing the plan to the device profile 122 , Checking for applicability may include estimating a degree of adaptation to functions based on the device profile 122 and by controlling the adaptation of the functions within an applicable functional boundary based on the device profile 122 exhibit.

The plan is validated if both the availability and applicability of the feature settings are true (Block 504 ). The configuration component 118 leads the plan (block 506 In one embodiment, performing the plan includes automatically executing individual configuration commands.

If the plan is not valid (block 504 ), assigns the configuration component 118 Revert the plan and undo any changes made to the plan and send a notification (such as an error signal) to the configuration translation component 110 (Block 508 ). For example, if the current clock frequency is already set at a maximum clock rate, any schedule attempting to increase clock speed will be rejected. The undo might be a source of inaccuracy in the performance profile of the reconfigurable memory device in the event of races.

If execution of the plan is successful (block 510 ), one or more logical devices are created on the reconfigurable storage device by associating the plan (ie, executed configuration commands) with a storage location (Block 512 ). Finally, the configuration component updates 118 the device profile 122 (Block 514 ).

According to the exemplary embodiment, a physical storage device may include a plurality of logical devices 122 host, and any logical device 121 can be configured independently and have its own characteristics. Thus, the reconfigurable memory system of the exemplary embodiment generates QoS-aware logical devices 121 , The logical devices 121 can apply to different technologies such as space partitioning or space partitioning (like a traditional OS partition), multi-queue (like NMVe standard), multi-tenancy (like Cloud PaaS), virtualization (like SR-IOV) Standard) and others which are device specific.

For example, assume that the user requires two 100 GB storage spaces in a single device: one for read-intensive workloads and the other for write-intensive workloads. With a simple reconfigurable memory device based on the multi-queuing space partitioning technique, the user can create two partitions (for example, / dev / sda [1, 2]) on the device, and then create a partition (eg, / dev / sdal) with a queue is linked as a read logical device and the other partition (e.g., / dev / sda2) associated with the other queue as a logical device for writing. Then, the device can provide the obtained reading performance by the logical device for reading without performance degradation due to writing. A typical storage device may show the long latency with this scenario due to garbage collection due to write.

Referring again to 1 can the configuration guide 138 the device profile 122 as the feedback 130 received via (a) feedback channel (feedback channels) (not shown). To reduce the data transfer overhead, the actual information sent to the bus may vary. The configuration advisor 138 monitors the one or more logical drives across one or more physical storage devices and determines if all the user requests 123 are fulfilled. The configuration advisor 138 also notifies the configuration planner 136 to adjust the plan based on the device profile in case of SLA violation.

6 FIG. 11 is a flowchart illustrating an execution flow of the configuration advisor. FIG 138 illustrated according to one embodiment. The configuration advisor 138 waits for the configuration component 118 to the device profile 122 including runtime performance data 134 the logical devices 121 as a feedback 130 to collect and send (block 600 ). In one embodiment, the host 100 have one or more feedback channels (feedback channels) having a standard interface to provide an input / output trace used in the reconfigurable memory device 102 are transferred. In another embodiment, the feedback channel (s) may have a special interface to it to the configuration advisor 138 to enable the performance of the reconfigurable storage device 102 to monitor.

In response to receiving the device profile 122 monitors the configuration advisor 138 the performance of the reconfigurable storage device 102 by comparing the feedback 130 with the user requirements 123 (Block 602 ). If the configuration advisor 138 determines that the runtime performance data 134 none of the logical devices the user requirements 123 hurt (block 604 ), then is programming the reconfigurable storage device 102 completed or complete (block 606 ). If any of the logical devices 121 fails, the user requirements 123 to comply, the configuration guide notifies 138 the configuration planner 136 (For example, by providing the runtime performance data 134 corresponding to the logical device) to provide a new logical device configuration plan (block 608 ), where settings and characteristics of the storage device 102 be dynamically adjusted.

In one embodiment, if a violation is detected, the configuration advisor may 138 also the device profile 122 from the reconfigurable storage device 102 to obtain the most recent comprehensive performance data to generate a new plan, since the notification can only include the performance of logical devices of interest. In an embodiment, the reconfigurable memory device 102 the device profile 122 periodically or whenever the status of the runtime performance data 134 changes, actively report. For example, assuming a user request, all read requests must be serviced within X ms, the configuration advisor 138 instruct the storage device to record the service time of read operations or set a read flag if the service time is over X ms. Rather than waiting for the device profile 122 that of the reconfigurable storage device 102 can be sent to the host 100 also the device profile 122 from the reconfigurable storage device 102 request periodically or on a request basis.

In one embodiment, the configuration advisor 138 be activated by the following events: when new user requests 123 be transmitted; if existing user requirements 123 to be changed; when the runtime performance data 134 to be updated; whenever the runtime performance data 134 needed (due to a timeout); and / or whenever the reconfigurable storage device 102 the user requirements 123 can not meet.

7 Figure 13 is a block diagram illustrating another embodiment for a reconfigurable memory system. This embodiment illustrates that a single configuration translation component 700 not just multiple logical devices on a physical reconfigurable storage device 702 It can also manage several other physical reconfigurable storage devices 704 . 706 and 708 can manage, each of which can have multiple logical devices.

A method and system for a reconfigurable storage system has been disclosed. The present invention has been described in accordance with the illustrated embodiments, and there may be variations for the embodiments, and any variation would be within the spirit and scope of the present invention. For example, the exemplary embodiment may be implemented using hardware, software, a computer-readable medium containing program instructions, or a combination thereof. Software written in accordance with the present invention must be stored in either a form of computer readable medium such as a memory, a hard disk, or a CD / DVD-ROM and must be executed by a processor. Accordingly, many modifications may be made by one skilled in the art without departing from the spirit and scope of the appended claims.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• KR 10-2012-0005764 [0001]

Claims (41)

A method of reconfiguring a memory system, the method being performed by at least one software component residing on at least one processor ( 114 ), comprising: receiving user request information ( 123 ) for a storage device ( 102 ) and automatically generating function settings for the storage device ( 102 ) from the user request information ( 123 ) and a device profile ( 122 ) for the storage device ( 102 ); and using the function settings to automatically save the storage device ( 102 ) into one or more logical devices ( 121 ), which have independent behavioral characteristics. The method of claim 1, further comprising: receiving a response ( 130 ) from the storage device ( 102 ) concerning runtime performance data ( 134 ) at least one of the storage device ( 102 ) and the logical devices ( 121 ); and responding to a determination that the runtime performance data ( 134 ) the user request information ( 123 ) dynamically adjust the configuration of one or more logical devices ( 121 ), the user request information ( 123 ). The method of claim 1, wherein the user request information ( 123 ) a high-level description of behavioral characteristics of the memory device ( 102 ) exhibit. The method of claim 2, wherein the high level description comprises at least one of a quality of service agreement (SLA) and quality of service (QoS) requirements. The method of claim 3, wherein the SLA further comprises specifications for any combination of attributes including capacity, addressing mode, protection, and performance. The method of claim 1, further comprising: using the device profile to determine whether a specification of the user request information ( 123 ) by the storage device ( 102 ) can be fulfilled. The method of claim 6, wherein the device profile ( 122 ) Performance models ( 132 ), which functions of the memory device ( 102 ), which have a correlation with the device performance, are assigned. The method of claim 6, further comprising: storing the device profile ( 122 ) on the storage device ( 102 ) and providing the device profile ( 122 ) for a host ( 100 ). The method of claim 6, further comprising: storing the device profile ( 122 ) in at least one of: a host ( 100 ) and a remote location hosted by the host ( 100 ) or the storage device ( 102 ) can be accessed. The method of claim 1, wherein automatically generating function settings further comprises: using the function settings to perform certain functions of the memory device ( 102 ) and to specify values for certain functions, wherein an effect of the functions global for all logical devices ( 121 ) or locally for a particular logical device ( 121 ). The method of claim 2, wherein the runtime performance data ( 134 ) a part of the device profile ( 122 ) and performance indices that are available via a self-monitoring, analysis and reporting technology (SMART) interface and vendor-specific runtime performance metrics that provide a minimum / maximum / average read / write latency, a minimum / maximum / average queue length, a minimum / maximum / average internal IOPS, a read / write gain ratio, and a garbage collection time. The method of claim 11, wherein the runtime performance data ( 134 ) further comprise: a device status, a device background job status, and a host background job status. A reconfigurable storage system comprising: a configuration translation component ( 110 ), which are hosted on a host ( 100 ), which user request information ( 123 ) for a storage device ( 102 ) and automatically receives function settings for the storage device ( 102 ) from the user request information ( 123 ) and a device profile ( 122 ) for the storage device ( 102 ) generated; and a configuration component residing on the storage device ( 102 ) executing the function settings from the host ( 100 ) and uses the function settings to automatically save the storage device ( 102 ) into one or more logical devices ( 121 ), which have independent behavioral characteristics. The system of claim 13, wherein the configuration translation component ( 110 ) is still configured to: 130 ) from the storage device ( 102 ) concerning runtime performance data ( 134 ) at least one of the storage devices ( 102 ) and the logical devices ( 121 ) to recieve; and in response to a determination that the runtime performance data ( 134 ) the user request information ( 123 ), the configuration of the one or more logical devices ( 121 ) dynamically adjust the user request information ( 123 ). The system of claim 13, wherein the user request information ( 123 ) a high-level description of behavioral characteristics of the memory device ( 102 ) exhibit. The system of claim 15, wherein the high level description comprises at least one of a quality of service agreement (SLA) and quality of service (QoS) requirements. The system of claim 16, wherein the SLA further comprises specifications for any combination of attributes including capacity, addressing mode, protection, and performance. The system of claim 13, wherein the configuration translation component ( 110 ) is further configured to change the device profile ( 122 ) to determine whether a specification of the user application information by the storage device ( 102 ) can be fulfilled. The system of claim 18, wherein the device profile ( 122 ) Performance models ( 132 ), which functions of the memory device ( 102 ), which have a correlation with the device performance, are assigned. The system of claim 18, wherein the device profile ( 122 ) on the ( 102 ) and the host ( 100 ) is made available. The system of claim 18, wherein the device profile ( 122 ) in at least one of the host ( 100 ) and a remote location hosted by the host ( 100 ) or the storage device ( 102 ) is accessible. The system of claim 13, wherein the configuration translation component ( 110 ) is further configured to use the function settings to perform certain functions of the memory device ( 102 ) and to specify values for certain functions, wherein an effect of the functions global for all logical devices ( 121 ) or local to a particular logical device ( 121 ). The system of claim 22, wherein the runtime performance data ( 134 ) a part of the device profile ( 122 ) and performance indices that are available via a self-monitoring, analysis and reporting technology (SMART) interface and provider-specific runtime performance metrics that have a minimum / maximum / average read / write latency, a minimum / maximum / average queue length, a minimum / maximum / average internal IOPS, a read / write gain ratio, and a garbage collection time. The system of claim 23, wherein said runtime performance data ( 134 ) further comprise: a device status, a device background job status, and a host background job status. An executable software product stored on a computer-readable storage medium having program instructions for reconfiguring a storage system, the program instructions being: receiving user request information ( 123 ) for a storage device ( 102 ) and automatically generating function settings for the storage device ( 102 ) from the user request information ( 123 ) and a device profile ( 122 ) for the storage device ( 102 ); and using the function settings to save the storage device ( 102 ) automatically into one or more logical devices ( 121 ), which have independent behavioral characteristics. The executable software product of claim 25, further comprising program instructions for: receiving a response ( 130 ) from the storage device ( 102 ) concerning runtime performance data ( 134 ) at least one of the storage devices ( 102 ) and the logical devices ( 121 ); and in response to a determination that the runtime performance data ( 134 ) the user request information ( 123 ), dynamically adjusting the configuration of the one or more logical devices ( 121 ), the user request information ( 123 ). An executable software product according to claim 25, wherein the user request information ( 123 ) a high-level description of behavioral characteristics of the memory device ( 102 ) exhibit. The executable software product of claim 27, wherein the high level description comprises at least one of a quality of service agreement (SLA) and quality of service (QoS) requirements. The executable software product of claim 28, wherein the SLA further comprises specifications for any combination of attributes including capacity, addressing mode, protection, and performance. The executable software product of claim 25, further comprising program instructions for: using the device profile ( 122 ) to determine if a specification of the user request information ( 123 ) by the storage device ( 102 ) can be fulfilled. An executable software product according to claim 30, wherein the device profile ( 122 ) Performance models ( 132 ), which functions of the memory device ( 102 ) having a correlation with the device performance. The executable software product of claim 30, further comprising program instructions for: storing the device profile ( 122 ) on the storage device ( 102 ) and providing the device profile ( 122 ) for a host ( 100 ). The executable software product of claim 30, further comprising program instructions for: storing the device profile ( 122 ) in at least one of: a host ( 100 ) and a remote location hosted by the host ( 100 ) or the storage device ( 102 ) can be accessed. The executable software product of claim 25, wherein automatically generating function settings further comprises program instructions for: using the function settings to perform certain functions of the memory device ( 102 ) and to specify values for certain functions, wherein an effect of the functions global for all logical devices ( 121 ) or locally for a specific logical device ( 121 ) is / are. An executable software product according to claim 34, wherein said runtime performance data ( 134 ) a part of the device profile ( 122 ) and performance indices that are available via a self-monitoring, analysis and reporting technology (SMART) interface and vendor-specific runtime performance metrics that provide a minimum / maximum / average read / write latency, a minimum / maximum / average queue length, a minimum / maximum / average internal IOPS, a read / write gain ratio, and a garbage collection time. An executable software product according to claim 35, wherein said runtime performance data ( 134 ) further comprise: a device status, a device background job status, and a host background job status. Host ( 100 ) having a storage device ( 102 ) reconfigured, comprising: a memory ( 106 ); a processor ( 114 ), which with the memory ( 106 ) is coupled; and a first software component generated by the processor ( 114 ) configured to: user request information ( 123 ) having quality of service agreement (SLA) specifications; the SLA specifications in feature settings based on a device profile ( 122 ) to transform; to determine optimal combinations of function settings; create multiple schedules containing configuration commands from the optimal combinations; and one of the plans taking into account the performance and status of the storage device ( 102 ) and to select the user requirements; and a second software component generated by the processor ( 114 ) configured to: receive the selected schedule; the plan to the reconfigurable memory ( 106 ) via a configuration channel; a response from the storage device ( 102 ) to recieve; and the plans in case of a negative response from the storage device ( 102 ) to iterate; and a third software component provided by the processor ( 114 ) which is configured to: the device profile ( 122 ) and the device profile ( 122 ) to the configuration planner ( 136 ) to send; the one or more logical device (s) ( 121 ) stored on the storage device ( 102 ) is generated, and to determine if all user requests are met; and the configuration planner ( 136 ) to update the plan based on the device profile ( 122 ) in case of SLA violation. Host ( 100 ) according to claim 37, further comprising: a plurality of performance models ( 132 ), which corresponds to each function of the reconfigurable memory device ( 102 ) correlations and anti-correlations between functions and scheduling algorithms, which are the performance models ( 132 ) and generate the configuration plans. Reconfigurable storage device ( 102 ), comprising: a main memory ( 112 ); a processor ( 114 ), which is connected to the main memory ( 112 ) is coupled; and a software component generated by the processor ( 114 ) configured to: receive a schedule having at least one configuration command containing function settings; availability and applicability of function settings in the plan based on a device profile ( 122 ) to check; if the plan is not valid, reject the plan and undo any changes made by the plan to validate the plan if both the availability and applicability of the feature settings are true; execute the plan and the device profile ( 122 ) to update; and one or more logical devices ( 121 ) on the reconfigurable storage device ( 102 ) by associating the executed plan with a storage location. Reconfigurable storage device ( 102 ) according to claim 39, wherein the software component checks the availability of the function settings by: determining whether the reconfigurable memory device ( 102 ) supports the function settings by comparing the plan with the device profile ( 122 ). Reconfigurable storage device ( 102 ) according to claim 40, wherein the software component checks the applicability of the function settings by: estimating a degree of adaptation Functions based on the device profile ( 122 ) and by controlling the adaptation of the functions within an applicable limit function based on the device profile ( 122 ).

Images