Elements of successful industrial-grade SSD design

By Scott Philips

VP Marketing

Virtium

January 19, 2015

Elements of successful industrial-grade SSD design

Editor's note: This is a follow-on to the blog, "Industrial-grade flash is important to embedded designs," written by Steve Larrivee of Cactus Technol...

Editor’s note: This is a follow-on to the blog, “Industrial-grade flash is important to embedded designs,” written by Steve Larrivee of Cactus Technologies. Going beyond just the flash media used, there are many more important considerations in the selection of industrial-grade SSDs.

While many storage suppliers tout their embedded SSDs as “industrial grade,” the truth is that very few actually are. Many are originally built as consumer-grade devices and repackaged, yet lack essential features. To truly be considered industrial-grade, SSDs must:

  • Be SLC NAND flash-based – SLC NAND is rated much higher in endurance and reliability than MLC NAND
  • Offer advanced flash management – It must enable more robust operation with higher endurance, data integrity, and data retention
  • Integrate power-fail corruption mitigation technologies – This is essential for embedded systems that operate 24/7/365
  • Backed by extensive design and reliability testing – This is required to adequately support hundreds of different applications, and vital in uncovering inherent flash media weaknesses that can be overcome with a suppliers’ strong design experience

Simply designing to an industry standard (e.g., MO-297) and using off-the-shelf industrial-rated components aren’t enough to meet the requirements of extreme or critical field applications. These drives would fail qualification, or worse, fail when deployed.

SSD features are just the tip of the iceberg. Selecting an industrial-grade SSD supplier highlights the need to evaluate multiple capabilities. Suppliers must have solid and well-tested hardware designs that are suited for high-volume manufacturing. They also need a thorough understanding of controller design, and for the highest reliability, ideally develop their own proprietary firmware.

Suppliers building their own firmware from the ground up have full control to provide the following advantages:

  • Firmware fine-tuning and fixes – Easier and faster firmware changes are possible, rather than waiting for the outside firmware developer
  • More efficient product design – This is accomplished with thorough test programs that reach deeper into the memory array
  • Advanced flash management feature implementation – This allows direct access to a knowledgeable supplier who can integrate drive performance enhancement features
  • Customer-specific design – Many customers still need support for legacy systems, which frequently have unique timings and power requirements
  • Non-traditional operating system support – In-house developed solutions, real-time operating systems, and Linux variants can be supported giving customers broader support for their applications

It’s important that proprietary firmware initially be “tuned” with an SSD’s flash to ensure that it’ll function reliably over the drive’s life. Each flash media type has its own unique program and erase characteristics that can be affected by timing, voltage, etc. Using “cookie cutter” firmware simply won’t recognize the intricacies of the flash and won’t be as efficient as one specifically built for it. Firmware also requires constant tweaks and fixes after the initial tuning. Similar to the design of a car engine, self-testing programs act as the mechanics of the drive and help ensure long-term functionality and reliability.

Thorough screening of an SSD during the manufacturing process such as scanning the flash media for bad blocks that don’t meet a set threshold is imperative. This is where strong, strategic relationships with leading flash media manufacturers are valuable to fully understand affected flash media nuances, and to gain access to specific modes and programming options required to conduct thorough testing.

Much planning goes into the choice of components from quality, product longevity, and replacement perspectives. To support industrial customers that require long product lifecycles, duty ratings, and tolerances figure highly in the selection of SSD components to avoid frequent re-qualifications. So does a comprehensive understanding of the various NAND types and knowledge of specific modes and other intricacies within the flash. NAND manufacturers typically provide these intricacies to select suppliers with which they have strategic relationships. Plus, suppliers that deliver proven, complete industrial-grade SSD solutions offer a locked bill of material (BOM) and support rigorous PCN processes.

When evaluating storage solutions for an embedded design, all of the considerations listed here are important. OEMs may find short-term cost disadvantages, but will realize long-term benefits from selecting a true industrial-grade SSD from a knowledgeable and experienced supplier partner.

Scott Phillips is in charge of executing Virtium’s business development strategies. Phillips has more than 20 years of experience working with many of the industry’s largest, multinational technology OEMs, and most recently managed the embedded/industrial solid-state storage solutions at sTec/HGST. He holds a Master’s in Business Administration.

Scott Phillips, Virtium

Professional experience includes 20+ years of marketing, product/brand management, and business development in the high-tech industry. Have worked for large multinational companies as well as small companies with responsibility for domestic and overseas accounts, primarily in the OEM space.

More from Scott

Categories
Storage