For more than a year, the SNIA Solid State Storage Technical Working Group working closely with the SNIA Solid State Storage Initiative (SSSI), has been hard at work developing an important new standard: the Solid State Storage Performance Test Specification (SSS PTS). The SSS PTS provides to all industry and end user stakeholders a standard methodology and nomenclature for measuring the performance of solid state storage devices. Many companies have been involved in the creation of this first iteration of the SSS PTS and contributed their expertise from all parts of the SSS industry, including the designers and manufacturers of SSS devices, computer systems, and controller chips, as well as end users. This broad base of participation ensures that the SSS PTS represents many points of view, not just one company or one segment of the industry.
SSS devices have been around for decades, but the market for them was small until recently when NAND Flash SSS began appearing that was reliable enough for both consumer and enterprise use and dramatically less expensive than previous SSS products and solutions. Greater market interest and adoption has resulted. But unlike mechanical spinning hard disk drives (HDD) where the performance of the basic medium is relatively predictable and managing it fairly straightforward, NAND Flash as a storage medium brings plenty of challenges; mitigating these has kept battalions of storage engineers well employed for years. The current crop of Flash engineering solutions has led to a broad and confusing array of Flash SSS products flooding the marketplace, with just as many confusing and often conflicting claims from SSS vendors about the basic performance of their products.
Today, there is no standard test methodology or specified test environment for measuring SSS device performance. As a result, each SSS manufacturer utilizes different measurement methodologies to derive performance specifications for their SSS products. When you read marketing collateral from different SSS vendors, you may not know if the performance numbers you see are burst or sustained performance, 100% read workload or a more realistic read/write mix, small data block transfers or large, etc. This makes comparing the performance of one SSS device against another essentially impossible. To address current SSS market confusion, the SNIA SSS TWG, working closely with SSSI, began developing the SSS PTS. This specification defines a suite of tests and test methodologies that effectively measure the performance characteristics of individual solid state storage products. When executed in a specific hardware/software environment, the SSS PTS will provide measurements of performance that may be fairly compared to those of other SSS products measured in the same way in the same environment.
The current SSS PTS version is focused on measuring NAND Flash based SSS at the device level, with an emphasis on defining nomenclature, metrics, and methodologies for performance test measurement. The SSS PTS discusses test platform and test tool requirements (to minimize hardware and software bottlenecks), recommends test practices (device purge, preconditioning, and measurement in steady state), defines basic client and enterprise device tests (IOPS, bandwidth, and latency), and sets forth standardized reporting formats. The SSS PTS is intended for use by individuals and companies engaged in the design, development, qualification, manufacture, test, acceptance, and failure analysis of SSS devices, systems, and subsystems. It is the intention of SNIA SSS TWG and SSSI to continue to develop and promote the SSS PTS enabling client and enterprise SSS industry and end user stakeholders to make realistic, accurate, and fair judgments about the performance of SSS devices.
The current version of the SSS PTS will be available in July, 2010 for public review. All interested parties are encouraged to download the work-in-progress SSS PTS draft at http://www.snia.org/publicreview/ and provide feedback at http://www.snia.org/feedback/. Comments coming out of the public review will be used to improve the specification. Status on updates to the specification and related white papers can also be found at http://www.snia.org/forums/sssi
The Solid State Storage Performance Test Specification brings clarity to an otherwise confusing marketplace, which will lead to greater confidence in solid state storage solutions and ultimately higher adoption rates of this very beneficial technology.
 ShareThis
Tags: BC/DR |
