Hybrid Storage System and Storage Method Based on Solid State Disk Caching 

The present invention relates to the field of storage systems, and more particularly to a storage system and a storage method for optimizing storage of a hybrid solid-state disk (SSD) and a mechanical disk (HDD).

The system comprises a mechanical magnetic disk, wherein storage space is divided into a solid state disk of multiple cache blocks as well as a request classification module, a request assignment device and a cache management device; and the cache management device comprises a request management unit.

Solution

• The invention discloses a hybrid storage system and a storage method based on solid state disk caching.
• According to the invention, the most suitable request is selectively cached on the solid state disk.
• The solid state disk is used to enhance overall performance of the system; the service life of the solid state disk is prolonged.
• The hybrid storage system cached based on solid-state disk and storage method that the present invention is provided, optionally will be most suitable request be buffered on solid-state disk, both using solid-state disk come lift system overall performance, at the same extend solid-state disk use time.
• Solid-state disk is compared, systematic function can be improved and can be reduced in large quantities again. The writing data quantity of mechanical disk is reached, so as to improve the service life of solid-state disk.

Application

The method aims to achieve solid-state disk performance at the cost level of mechanical disk drives. It seeks to strike an optimal balance between performance and cost, offering users transparent high performance, large storage capacity, and block-level devices with low power consumption.

Comment

• The solid state disk is used to enhance overall performance of the system; the service life of the solid state disk is prolonged.
• A compromise tradeoff point is selected between performance and life-span of a solid state disk.

Solid-State Drive Management and Control

The present invention uses the device interface and the SSD interface that are configured to transmit commands for setting operating parameters for the SSD. Example commands can include set recovery time, set retention period, set wear leveling, and/or other suitable commands. For example, the commands set recovery time and set retention period may set at least one of a maximum value, minimum value, average value, or allowed range for cell recovery time and retention period, respectively. The command set wear leveling may turn on (or off) dynamic or static wear leveling.

The SSD component is configured to receive operating parameters from the management engine and control operation of the SSD based on the received operating parameters. For example, in one embodiment, the SSD component can include a flash memory controller (“FMC”). The FMC receives operating parameters from the management engine via the SSD interface. Based on the received operating parameters, the FMC controls operations of accessing (e.g., reading, writing, and erasing), wear leveling (e.g., static, dynamic, or no wear leveling), block picking, garbage collection, and/or other suitable operations of the SSD.

Solution

• The present technology is directed to management and/or control of operational characteristics of SSDs at an operating system and/or application layer.
• In one aspect, the present technology provides a management engine as a part of an operating system and/or application for setting operating parameters of SSDs based on a target operation profile.
• For example, the management engine can include an input interface configured to accept input of endurance, retention, performance, and/or other target operation profiles from an operator and/or application.
• The management engine can also include a process component configured to determine one or more operating parameters for SSDs based on the received input and predetermined operating policies.
• Example operating parameters can include cell recovery time, retention period, wear leveling, and/or other suitable operating parameters.

Application

The method helps in recognizing that different computing tasks assigned to a computing device may require different tradeoffs of SSDs coupled to the computing device as persistent storage. For instance, time sensitive tasks (e.g., web search) may require high performance, while other tasks (e.g., data backup) may require high endurance.

Comment

• The disclosed method helps in managing the foregoing tradeoffs of SSDs based on computing task demands, and thus prolonging usable life of SSDs.
• Further, reducing costs of using SSDs as storage devices, and improving operation performance of computing devices when compared to conventional techniques.

Scrubber Driven Wear Leveling in out of Place Media Translation

Aspects of the present disclosure are directed to wear-leveling process implemented in a memory subsystem. A memory subsystem is also hereinafter referred to as a “memory device.”

The embodiments include a wear-leveling manager that provides an efficient wear-leveling process. The wear-leveling manager can operate in a memory subsystem that utilizes an out-of-place translation layer such as a flash translation layer. The wear-leveling manager can be implemented in or include a media management agent and/or garbage collection related processes. In some embodiments, the wear-leveling manager identifies low write count chunks (e.g., managed units) of the memory subsystem. The wear-leveling manager invalidates the low write count chunks. The wear-leveling manager can queue these invalidated low write count chunks to be utilized for write requests thereby distributing wear to these low write count chunks.

Solution

• The present disclosure generally relates to memory subsystem management, and more specifically, relates to a wear-leveling scheme for memory subsystems.
• The embodiments include a wear-leveling manager that provides an efficient wear-leveling process. The wear-leveling manager can be implemented in or include a media management agent and/or garbage collection related processes.
• The wear-leveling manager identifies low write count chunks of the memory subsystem. The wear-leveling manager invalidates the low write count chunks. The wear-leveling manager can queue these invalidated low write count chunks to be utilized for write requests thereby distributing wear to these low write count chunks.
• A cartridge manager, which can be a part of the wear-leveling manager or a separate component, can further organize invalidated chunks into groupings referred to as ‘cartridges.’

Application

The method provides benefits such as defect management, power consumption management, a-priori channel conflict resolution, write amplification        reduction, proximity disturb management, and performance defect management is improved by enabling regular offline (e.g., unmapped) media health assessment using the tracked write counts, which enables graceful capacity degradation by retiring bad chunks (i.e., high write count chunks).

Comment

• The memory subsystem includes a wear-leveling manager that can increase the operational life of the memory components.
• The method disclosed is responsible for operations such as wear leveling operations, garbage collection operations, and caching operations that are associated with the memory components.

Enhanced Garbage Collection 

The present invention discloses example method of enhanced garbage collection operations, in accordance with certain embodiments of the present disclosure. Method provides an example embodiment of a process for sorting data based on data streams to improve device performance.

Some solid state memory devices have a limited lifespan, often measured in a number of program-erase (PE) cycles that portions of the physical memory can undergo before the memory becomes unreliable at retaining data. Further, performing write operations on solid state memory may be relatively slow compared to performing read operations. Therefore it may be advantageous to reduce write amplification (WA) at solid state memories. WA can describe an undesirable phenomenon associated with solid-state drives (SSDs) where the actual amount of physical information written to nonvolatile memory is a multiple of the logical amount intended to be written. Operations such as wear leveling and garbage collection can involve moving and rewriting user data and metadata more than once, resulting in a higher WA, slowing drive performance and putting wear on the memory.

Solution

• A DSD (data storage device) may include a garbage collection module (GCM). The GCM may perform the methods and processes described herein to select data and blocks for collection, and to determine how to organize and rewrite the valid data collected during garbage collection.
• For example, the GCM may analyze valid data stored to blocks, and evaluate various attributes of the valid data to determine which blocks to garbage collect.
• The GCM may also determine whether and how to reorganize the collected data to improve data access efficiency.
• Additionally, the GCM may determine whether and how much to throttle processing on host commands while the garbage collection operations are performed in order to limit variability in host command throughput.

Application

The solution is helps in sorting data for storage based on the data stream via which the data was received which can improve access efficiency and reduce write amplification at the device.

Comment

• The method is suitable for prolonging the lifespan of solid state memory devices.
• The method further provides improving device responsiveness and efficiency, and reduce write amplification.

Using Recurring Write Quotas to Optimize Utilization of Solid State Storage in a Hybrid Storage Array 

The invention provides operation balances wearing of the SSD storage components, and enables the SSD storage components to reach their optimum lifespan. That is, the storage processing circuitry uses the recurring write quotas to pace (or regulate) usage of the SSD storage components so the SSD storage components are not overly worn prematurely. Without such operation, utilization of the SSD storage components may become grossly uneven and some SSD storage components could become overused and require replacement too soon (i.e., ahead of schedule).

A storage processing circuitry monitors the operation of each SSD storage component and may adjust the write quotas based on healthiness indicators. That is, due to variations in each SSD storage component, some SSD storage components may provide higher than average errors (write faults, read faults, etc.), average errors, below average errors, and so on. For each SSD storage component that exhibits higher than average errors, the storage processing circuitry may reduce the respective recurring write quota since that SSD storage component could prematurely fail. Additionally, for each SSD storage component that exhibits an average number of errors, the storage processing circuitry may maintain the respective recurring write quota as initially set or perhaps make minor adjustments based on factors such as the actual number of PE-cycles used by that SSD storage component. Furthermore, for each SSD storage component that exhibits lower than average errors, the storage processing circuitry may increase the respective recurring write quota since that SSD storage component may exhibit a life significantly higher than what is even prescribed by the SSD manufacturer.

Solution

• The present invention discloses an electronic circuitry configured to reset the recurring write quotas and repeat this process routinely (e.g., daily) thus enabling the data storage equipment to more evenly wear the SSD storage components.
• Any alerts to the operator enables the operator to modify how the SSD storage components to improve the ability of the SSD storage components to safely reach the planned lifetime before reaching the PE-cycle limits.
• The electronic circuitry derives the recurring write quota for each SSD storage component based on a variety of reliability and healthiness factors.
• The electronic circuitry resets or re-initializes the recurring write quota for each SSD storage component to the full recurring write quota. At this time, if writes had been blocked to certain SSD storage components because their recurring write quotas had been fully consumed during the earlier time period, the writes may now be unblocked thus again re-enabling writes to those SSD storage components.

Application

The solution improves write amplification of the SSD storage(s) based on a variety of reliability and healthiness factors.

Comment

• The solution is providing a balance between the goals of minimizing latency (e.g., as data is written to SSD storage) and reaching an expected SSD lifetime (e.g., as further data is written to HDD storage to preserve SSD lifetime)..
• Further, the write cycles may be increased because of increase in the life expectancy of the SSD storage(s).

About Effectual

Effectual Services is an award-winning Intellectual Property (IP) management advisory & consulting firm offering IP intelligence to Fortune 500 companies, law firms, research institutes and universities, and venture capital firms/PE firms, globally.

Through research & intelligence we help our clients in taking critical business decisions backed with credible data sources, which in turn helps them achieve their organisational goals, foster innovation and achieve milestones within timelines while optimising costs.