Fighting Downtime: Low-Latency SSDs Improve Storage Redundancy

Samsung Insights | Jul 10, 2019 at 2:00 PM

Data is one of the most critical assets of any business, and most organizations require 24/7 access. Even a system utilizing multiple low-latency solid state drives (SSDs) providing incredible performance can experience hardware failures with the potential to cause downtime or data loss, leading to decreased productivity, revenues and customer satisfaction.

This means protecting that data and the access to it is essential to the bottom line. Since there are multiple components in the storage subsystem, including drives, a storage controller, a motherboard and more, a storage system needs to survive the failure of one or more of those components. Many modern motherboards include several SATA controller channels, allowing for four, five, six or eight channels, each with one SATA SSD, which can then be combined into a single redundant array of independent disks (RAID) volume. RAID controllers can also be used in PCIe slots. Hardware RAID is generally faster than the software RAID used with multiple motherboard channels.

RAID disks (or inexpensive disks) were developed to help overcome failures of hard disk drives (HDDs). Common versions include RAID 0, 1, 0+1, 5 and 6. Some offer higher performance than a single drive, while others are geared more towards redundancy, and all combine multiple drives into a single virtual volume.

Exploring Redundancy

Samsung products throughout their enterprise SSD line have features specifically intended to improve resilience and data protection.

For example, the Samsung 883 DCT, 983 DCT, and 983 ZET built-in capacitors that enable the system to finish writing data to the drive even if the power fails. Additionally, Samsung SSDs in the enterprise feature higher reliability than consumer-grade drives as well as end-to-end data protection and sophisticated error-correction algorithms.

Low latency, high throughput and low random access times all improve rebuild times for RAID sets, ensuring that volumes are back in a redundant configuration as quickly as possible. Picking the right RAID volume is also essential. While a five-drive RAID5 volume might take hours to rebuild after a drive failure, a 0+1 RAID volume with low-latency, high-speed SSDs could rebuild in minutes.

SSD Lifetime Advantage

Another potential issue is the overall lifetime of the SSD. Unlike HDDs, which most commonly fail due to mechanical problems, SSDs are entirely electronic, and generally more reliable to begin with. One of the most common failures with SSDs is due to the wearing of individual cells. Each cell in an SSD has a limited lifespan. Although modern SSDs last far longer than early ones, buyers should pay attention to the drive writes per day (DWPD). This number estimates how many times you can completely rewrite the drive for each day of its warranty period.

Consumer-grade SSDs may be as low as 0.1 DWPD, while enterprise-class SSDs like the 983 ZET offer up to 10 DWPD. Storing critical data on more reliable drives will contribute to overall system reliability.

Keeping data secure requires a multifaceted approach: backups, mirroring of volumes, replication of data in other data centers or in the cloud. Each of these aspects depends on writing the original source data correctly and storing it securely. Investing in the most reliable SSDs available is a great place to start, but many IT administrators still prefer to also implement RAID to optimize their data storage.

Download your free guide to improving the lifespan of your SSDs by over-provisioning. Looking to enhance storage performance in your data center? Take a brief video walkthrough of Samsung’s enterprise SSD lineup.