Solid-State Drive (SSD) Requirements and Endurance Test Method
JEDEC SOLID STATE TECHNOLOGY ASSOCIATION
NOTICE JEDEC standards and publications contain material that has been prepared, reviewed, and approved through the JEDEC Board of Directors level and subsequently reviewed and approved by the JEDEC legal counsel. JEDEC standards and publications are designed to serve the public interest through eliminating misunderstandings between manufacturers and purchasers, facilitating interchangeability and improvement of products, and assisting the purchaser in selecting and obtaining with minimum delay the proper product for use by those other than JEDEC members, whether the standard is to be used either domestically or internationally. JEDEC standards and publications are adopted without regard to whether or not their adoption may involve patents or articles, materials, or processes. By such action JEDEC does not assume any liability to any patent owner, nor does it assume any obligation whatever to parties adopting the JEDEC standards or publications. The information included in JEDEC standards and publications represents a sound approach to product specification and application, principally from the solid state device manufacturer viewpoint. Within the JEDEC organization there are procedures whereby a JEDEC standard or publication may be further processed and ultimately become an ANSI standard. No claims to be in conformance with this standard may be made unless all requirements stated in the standard are met. Inquiries, comments, and suggestions relative to the content of this JEDEC standard or publication should be addressed to JEDEC at the address below, or call (703) 907-7559 or www.jedec.org Published by JEDEC Solid State Technology Association 2010 3103 North 10th Street Suite 240 South Arlington, VA 22201-2107 This document may be downloaded free of charge; however JEDEC retains the copyright on this material. By downloading this file the individual agrees not to charge for or resell the resulting material. PRICE: Contact JEDEC
Printed in the U.S.A. All rights reserved
DON??T VIOLATE THE LAW!
This document is copyrighted by JEDEC and may not be reproduced without permission. Organizations may obtain permission to reproduce a limited number of copies through entering into a license agreement.
For information, contact:
JEDEC Solid State Technology Association 3103 North 10th Street Suite 240 South Arlington, VA 22201-2107 or call (703) 907-7559
JEDEC Standard No. 218 Page 1
SOLID STATE DRIVE (SSD) REQUIREMENTS AND ENDURANCE TEST METHOD
(From JEDEC Board Ballot JCB-10-53, formulated under the cognizance of the JC-64.8 Subcommittee on Solid State Drives.)
This standard defines JEDEC requirements for solid state drives. For each defined class of solid state drive, the standard defines the conditions of use and the corresponding endurance verification requirements. Although endurance is to be rated based upon the standard conditions of use for the class, the standard also sets out requirements for possible additional use conditions as agreed to between manufacturer and purchaser. Qualification of a solid state drive involves many factors beyond endurance and retention, so such qualification is beyond the scope of this standard, but this standard is sufficient for the endurance and retention part of a drive qualification. This standard applies to individual products and also to qualification families as defined in this standard. The scope of this standard includes solid state drives based on solid-state non-volatile memory (NVM). NAND Flash memory is the most common form on memory used in solid state drives at the time of this writing, and this standard emphasizes certain features of NAND. The standard is also intended to apply to other forms of NVM.
The revision of the referenced documents shall be that which is in effect on the date of the qualification plan. JESD22-A117, Electrically Erasable Programmable ROM (EEPROM) Program/Erase Endurance and Data Retention Stress Test JESD47, Stress-Test-Driven Qualification of Integrated Circuits JEP122, Failure Mechanisms and Models for Semiconductor Devices JESD219, Solid State Devices (SSD) Endurance Workloads
Terms, definitions, abbreviations, and technical background Cycling pool
Erase blocks used by the SSD during read, write, or erase operations at a specific point in time.
NOTE The SSD may have additional erase blocks besides those in the current cycling pool that may be used as spares or for other purposes. The cycling pool is typically larger than the user-accessible LBA count.
JEDEC Standard No. 218 Page 2
Terms, definitions, abbreviations, and technical background (cont??d) Data Error
A type of failure in which the drive fails to return correct data to the host.
NOTE One data error occurs if a read of a logical sector causes the drive to return an unrecoverable error message or to return incorrect data. Data errors are counted as such even if they are transient. See 7.1.2 for further discussion of data errors.
The ability of the SSD to retain data over time. Synonymous with retention in this document. 3.4 Endurance
The ability of an SSD to withstand multiple data rewrites. 3.5 Endurance failure
A failure that is caused by endurance stressing.
NOTE 1 Endurance failures may be data error types or functional failure types, as described above.
NOTE 2 In an endurance stress, some failures may occur that are unrelated to endurance. For example, a solder joint could fail. A failure is considered unrelated to endurance if it was not caused by the endurance stress itself (i.e., if it was not caused by the repetitive writing of data to the drive). Such non-endurance failures are not considered as part of endurance verification. NOTE 3 A number of distinct failure mechanisms are responsible for endurance failures, and in general these are accelerated in different ways by temperature and other adjustable qualification parameters. For example, in floatinggate memories failure may be caused by charge trapping (normally accelerated by lower temperatures) in the charge transfer dielectric or by oxide rupturing (normally accelerated by higher temperatures) in the transfer dielectric or in peripheral dielectrics. For these reasons, endurance may depend on temperature but it is not known a priori whether high temperature is worse than low temperature, or vice-versa.
Endurance Rating (TBW rating)
The number of terabytes that may be written to the SSD such that the SSD meets the requirements defined in 6.2.
NOTE Several factors impact the endurance rating including how optimally the wear leveling has been implemented, write amplification factor and the cycling capability of the NAND components. The relationship between TBW, write amplification factor and the wear-leveling efficiency are highly dependent on the workload applied for the characterization of endurance.
The smallest addressable unit for erase operations, typically consisting of multiple pages.
JEDEC Standard No. 218 Page 3
Terms, definitions, abbreviations, and technical background (cont??d) Failure
The noncompliance of an SSD to the electrical or physical requirements specified for the device.
NOTE Failures may be permanent or transient. For the purpose of this standard, a permanent failure is an SSD that fails sometime during a reliability stress and continues to fail at the final test at the end of that same stress. A transient failure is an SSD that fails during a reliability stress but passes in the final test at the end of that stress.
A failed drive that fails to function properly in a way that is more severe than having a data error.
NOTE See 7.1.2 for discussion of functional failures.
Functional Failure Requirement (FFR)
The allowed cumulative functional failures over the TBW rating. See 6.3. 3.11 Gigabyte (GB)
Approximately equal to 109 bytes when used in reference to SSD capacity (see 4) 3.12 Host
The computer system, test system, or other device, which writes data to and reads data from the SSD. 3.13 Host writes
Data transmitted through the primary SSD interface to be written to the SSD. 3.14 Logical Block Address (LBA)
The logical address used to reference a data sector (block) in the drive.
NOTE 1 NOTE 2 LBA is synonymous with the data sector itself. Block in the drive is a logical construct separate from that of an erase block in the NVM.
Non-Volatile Memory (NVM)
A memory which retains data after the power is turned off.
NOTE The non-volatile memories considered in this standard are capable of being electrically rewritten.
JEDEC Standard No. 218 Page 4
A sub-unit of an erase block consisting of a number of bytes which
can be read from and written to in single operations, through the loading or unloading of a page buffer and the issuance of a program or read command. 3 Terms, definitions, abbreviations, and technical background (cont??d) 3.17 Program/erase cycle (p/e cycle)
The writing of data to one or more pages in an erase block and the erasure of that block, in either order. 3.18 Qualification Family
A group of SSD products which differ only in storage capacity and in minor design details that are directly related to the capacity differences.
NOTE To be considered part of the same qualification family, SSD products must use the same nonvolatile memory products, or different nonvolatile memory products that are themselves part of the same component qualification family (defined in JESD47). The SSD products must also use the same controller and the same firmware, except to the extent that the firmware requires different settings to support the different capacities of the drives. The SSD products must also have the same ratio of TBW specification to capacity; for example, a 100 GB drive with a 100 TBW specification could be in the same family as a 50 GB drive with a 50 TBW specification. Due to the complexity of SSD designs, it is beyond the scope of this specification to completely define what constitutes a qualification family. The burden of proof falls upon the SSD manufacturer who chooses to use the qualification family concept.
A data error occurring when the SSD is read after an extended period of time following the previous write.
NOTE A number of distinct failure mechanisms are responsible for retention failures, and in general these are accelerated in different ways by temperature and other adjustable qualification parameters. For example, in floatinggate memories, failure may occur due to defects that allow charge to leak through the transfer dielectric or by the detrapping of charge in the transfer dielectric; the former can be weakly accelerated or even decelerated by high temperature, and the latter can be highly temperature accelerated (see JEP122). For these reasons, retention may depend on temperature but it is not known a priori whether high temperature is worse than low temperature, or viceversa.
Solid State Drive (SSD)
A solid state drive (SSD) is a non-volatile storage device. A controller is included in the device with one or more solid state memory components. The device should use traditional hard disk drive (HDD) interfaces (protocol and physical) and form factors 3.21 Terabyte (TB)
For the purpose of this standard, a terabyte is equal to 11012 bytes.
JEDEC Standard No. 218 Page 5
Terms, definitions, abbreviations, and technical background (cont??d) Uncorrectable Bit Error Rate, or ratio (UBER)
A metric for the rate of occurrence of data errors, equal to the number of data errors per bits read. Mathematically,
number of data errors number of bits read
NOTE Although the UBER concept is in widespread use in the industry, there is considerable variation in interpretation. In this standard, the UBER values for SSDs are to be lifetime values for the entire population. The numerator is the total count of data errors detected over the full TBW rating for the population of SSDs, or the sample of SSDs in the endurance verification. A sector containing corrupted data is to be counted as one data error, even if it is read multiple times and each time fails to return correct data. The denominator is the number of bits written at the TBW rating limit, which aligns to the usual definition of errors per bit read when the read:write ratio is unity. See 7.1.1 for a further discussion of UBER calculation.
Methods employed by the drive to spread the p/e cycles across the NVM physical locations even when the workload may be unevenly distributed across the logical drive capacity. 3.24 Workload
The detailed sequence of host writes and reads (including data content and timing) applied to the drive during endurance testing. See JESD219 for specification of workload. 3.25 Write amplification factor (WAF)
The data written to the NVM divided by data written by the host to the SSD.
NOTE 1 For the purpose of calculating WAF, data written by the host is considered to be in multiples of drive capacity. For example, if 150GB of data is written to an SSD with capacity of 100GB, the data written by the host is considered to be 1.5. Also, data written to the NVM is considered to be the average number of p/e cycles experienced by NVM blocks in use in the SSD. For example, if the average number of p/e cycles is 3, then the data written to the NVM is considered to be 3. In this example, the WAF would be 2. NOTE 2 Write amplification factor will depend on the workload and may vary over the lifetime of the device.
JEDEC Standard No. 218 Page 6
Per the definition of SSD in this specification, the SSD may replace or co-exist with hard disk drives (HDD??s) in an application. To be consistent with the established IDEMA calculation for HDD capacity, SSD capacity shall be defined as user-addressable capacity as calculated using the following formula based on an LBA size of 512 bytes (LBA??s reserved for device use are not included in this calculation): User-addressable LBA count = 21168 + (1953504 x SSD Capacity in Gbytes); or SSD Capacity in Gbytes = (User-addressable LBA count - 21168) / 1953504. Due to the confusion in the industry and operating systems possibly using a binary calculation to report drive capacity, it is strongly recommended that product specifications clearly define how the SSD capacity is calculated.
SSD form factors include, but are not limited to, traditional HDD form factors. Many HDD form factors were developed in the SFF Committee. The SFF Committee web site, www.sffcommittee.org, is an excellent resource for finding mechanical specifications of drive form factors and the associated connector locations. Form factor publications that have become JEDEC, EIA, or other recognized standard organization specifications reference the standard number on the first page of the SFF publication.
Application classes and Endurance and Retention Rating Purpose of application classes
It is recognized that there are different levels of demand on the SSD based on different applications and that a different level of testing should be used to verify the SSD suitability for the particular application. The defining of classes provides a means of developing standard test methodologies that address multiple usage models by setting appropriate test limits based on the application class. For the purpose of endurance, this standard defines two application classes: client and enterprise. Section 6.3 defines the quantitative requirements for the application classes. These classes are not all-inclusive and it is understood that variations such as the operating system and application architecture make a significant impact to the workload of the SSD. These classes provide a means providing parameters for standardized endurance ratings so that the end user may use the endurance rating as a factor in determining if an SSD is suitable for his particular application.
JEDEC Standard No. 218 Page 7
Application classes and Endurance and Retention Rating (cont??d) SSD Endurance Rating
The SSD manufacturer shall establish an endurance rating for an SSD that represents the maximum number of terabytes that may be written by a host to the SSD, using the workload specified for the application class, such that the following conditions are satisfied: 1) 2) 3) 4) The SSD maintains its capacity The SSD maintains the required UBER for its application class The SSD meets the required functional failure requirement (FFR) for its application class The SSD retains data with power off for the required time for its application class
This rating is referred to as TBW. Requirements for UBER, FFR, and retention are defined in 6.3 for each application class. In addition to the standard endurance rating determined for the requirements in 6.3, the manufacturer may determine ratings for other sets of requirements as agreed to with purchasers. The methods described in this standard may be used to determine the endurance and retention verification procedures for those additional sets of requirements. See 7.3. 6.3 Requirements for standard classes of SSDs
This standard is based on a use scenario in which the SSDs are actively used for some period of time during which the SSDs are written to their endurance ratings, followed by a power-down time period in which data must be retained. Table 1 lists the requirements by SSD class. Table 1 ?ª SSD Classes and Requirements Workload Active Use Retention Use Functional (see JESD219) (power on) (power off) Failure Requirement (FFR) Client 40 oC 30 oC ?Ü3% 8 hrs/day 1 year Enterprise 55 oC 40 oC ?Ü3% 24hrs/day 3 months
UBER Requirement ?Ü10-15 ?Ü10-16
The endurance rating shall be determined for the workload and use temperatures required for the particular class of SSD. Since workloads are expected to change as applications evolve, workloads are described in a separate JEDEC standard (JESD219). The active-use and retention-use columns specify the temperatures and times for the two time periods of SSD use. The retention use temperature for client (30 oC) is also the temperature for the 16 hrs/day in which the SSD is off. Temperatures in Table 1 and elsewhere in this standard are SSD case temperatures, as reported by the drive if the drive has temperature reporting capability. The temperatures are intended to represent the relevant temperatures over the respective time periods, for the purpose of endurance and retention estimation, not the absolute max/min specifications to be found on the SSD datasheet. The retention use time is the period over which data must be retained with power off. The FFR and UBER columns specify the criteria for functional failures and UBER that must be met for drives written to their endurance rating limits and then being subjected to the specified retention time.
JEDEC Standard No. 218 Page 8
Application classes and Endurance and Retention Rating (cont??d) Estimation of the SSD endurance rating
An informative example of an SSD endurance rating estimation is provided in Annex A. The estimation method shown in Annex A is not suitable to general users of the drive because the method requires information about the drive design and data from the drive itself that will in general be available only to the manufacturer. The method in Annex A is also limited in scope to SSDs which use NAND components qualified according to JESD47 or, as described in 7.2.5, modified versions of the JESD47 qualification flow. The method in Annex A is not sufficient for verification that an SSD meets the endurance requirements. Verification of an endurance rating is defined in clause 7.
Endurance and Retention Stress Method
This section describes two approaches for endurance verification: a direct method (7.1) and a set of extrapolation methods (7.2). Both consist of endurance verification followed by retention verification. With the direct method, SSDs are written to their TBW ratings and put through retention testing as well. The direct method is to be followed if the full TBW rating can be reached in a 1000 hour stress. If this is not possible, then the extrapolation approach is acceptable. The manufacturer is allowed to use the direct method even if this takes longer than 1000 hours. If an SSD product from a qualification family has been qualified using this standard, then subsequent products need only data from a 1000-hour direct method evaluation, even if this results in those drives not being fully stressed to their TBW rating limits. 7.1 Direct method
SSDs are stressed to their stated endurance rating (in TBW) using specified workloads. The endurance stressing is to be performed at both high and low temperatures. Following this endurance stressing, retention testing shall be performed. Since the retention use time requirements (Table 1) are long, extrapolation or acceleration is required to validate that the SSD meets the retention requirement. Two approaches are acceptable for incorporating both high and low temperatures into the endurance stressing: the ramped-temperature approach and the split-flow approach. The approach used is up to the discretion of the manufacturer. Flowcharts summarizing the two approaches are shown in Figure 1. 7.1.1 Sample size and acceptance criteria
For the first product to be qualified in a qualification family, the sample shall consist of SSDs from at least three nonconsecutive
production lots and from all fabrication plants responsible for the manufacture of the nonvolatile memory used in the SSD. For subsequent products from a qualification family, a single production lot is sufficient. The number of SSDs in the sample shall be sufficient to establish that both the FFR and UBER requirements are met at 60% confidence.
JEDEC Standard No. 218 Page 9
Sample size and acceptance criteria (cont??d)
The sample sizes and acceptance criteria are defined by the following two equations, which mathematically embody the 60% requirement of the previous sentence:
UCL( functional _ failures) ?Ü FFR SS
UCL(data _ errors) ?Ü min (TBW , TBR ) 8 10 UBER SS
where functional_failures is the acceptable number of functional failures data_errors is the acceptable number of data errors min(x,y) is the minimum of x and y (see also 7.1.3) FFR and UBER are the drive requirements expressed as fractions TBW is the endurance rating in terabytes written TBR is the number of TB read (see 7.1.3) SS is the sample size in number of drives UCL() is an upper confidence limit function defined in Table 2 (see JESD47 for background) Table 2 ?ª Values of UCL(x)
AL 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 n 0.92 2.03 3.11 4.18 5.24 6.29 7.34 8.39 9.43 10.48 11.52 12.55 13.59 14.62 15.66 16.69 17.72 18.75 19.78 20.81 AL 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 n 21.84 22.87 23.89 24.92 25.94 26.97 28.00 29.02 30.04 31.07 32.09 33.12 34.14 35.16 36.18 37.20 38.22 39.24 40.26 41.29 AL 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 n 42.30 43.32 44.35 45.36 46.38 47.40 48.42 49.43 50.46 51.47 52.49 53.51 54.52 55.55 56.56 57.58 58.60 59.61 60.63 61.64 AL 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 n 62.66 63.68 64.69 65.71 66.72 67.74 68.75 69.77 70.79 71.80 72.82 73.83 74.85 75.86 76.88 77.89 78.91 79.92 80.94 81.95 AL 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 n 82.97 83.98 84.99 86.00 87.02 88.03 89.05 90.06 91.08 92.08 93.10 94.11 95.13 96.14 97.15 98.16 99.18 100.19 101.21 102.22
A common sampling plan is called an accept-on-zero plan because the evaluation will pass if there are zero failures and fail if there are greater than zero failures. For an accept-on-zero plan, the UCL value is 0.92.
NOTE Equations 2 and 3 represent two separate sample-size requirements and two separate acceptance criteria. In a particular