Choosing the Right SSDs for Your Data Center

By Doug Rollins, Principal Engineer, Systems Marketing, Enterprise SSDs, Micron Technology, Inc.

Solid state drives (SSDs) have made their way into the enterprise data center, and for good reason.  When compared to conventional hard drives, SSDs offer data center applications better performance, better responsiveness, lower power, and predictable longevity—all key attributes desired by data center and system managers.

Can an administrator or system designer use any SSD and expect great results?  Are there really any differences between enterprise-class and client-class SSDs, aside from price?

In this article we’ll take a look at some of the fundamental differences between enterprise-class and client-class SSDs to help you make the best choice.

Data Path Protection

The data path refers to the route data takes inside the SSD, starting at the drive interface, moving through the various buffers and queues, and finally ending up in the NAND Flash memory for long-term storage.  Data path protection is the means by which an SSD protects data as it moves inside the SSD.  In client-class SSDs, it is common practice to protect data by using parity and cyclic redundancy check (CRC).  However, true enterprise-class SSDs usually go one step further and add a memory protection mechanism—an extra layer of protection akin to a “belt and suspenders” approach.  Applying this additional protection to the memory in the data path helps guard against any data errors in the dynamic random access memory (DRAM) itself.

Power-Loss Protection

Data that has been written to the NAND in the SSD is called “data at rest.”  Data that is in the process of being written by the host computer is called “data in motion.”  Because these are live operations, the SSD has to be concerned with power loss. If power is lost while these (and other) operations are in progress, there is a potential for data loss.

In most client-class SSDs, only data at rest is protected against power loss; data in motion is not.  Enterprise-class SSDs, like Micron’s M500DC, employ additional hardware and firmware features to ensure that both data at rest and data in motion are protected against power loss.

Latency

Latency is an expression of how responsive an SSD is.  The lower the latency, the more responsive the SSD is, and the more responsive the application running on that SSD is.  Enterprise-class SSDs have a lower overall latency and a much smaller difference between the average and maximum latency, which enables enterprise applications to respond better and more consistently than client-class SSDs.

Performance

Although the read performance of enterprise- and client-class SSDs is very similar, the write performance tends to be quite different. Enterprise-class SSD write performance is much faster, especially with smaller, more random workloads as would be found in databases, virtualized environments, and other enterprise applications.  Enterprise-class SSDs excel with these workloads, but client-class SSDs struggle.

What’s the Best Choice?

While it’s possible to use client-class SSDs in the data center, enterprise-class SSDs tend to provide superior results due to their more robust data path protection, enhanced power-loss protection, lower latency, and superior write performance.

Doug Rollins, Principal Engineer, Systems Marketing, Enterprise SSDs, Micron Technology, Inc.,208-368-4000; email: dougrollins@micron.com; web: www.micron.com

Micron Technology, Inc., is a global leader in advanced semiconductor systems. Micron’s broad portfolio of high-performance memory technologies—including DRAM, NAND and NOR Flash—is the basis for solid state drives, modules, multichip packages, and other system solutions. Backed by more than 35 years of technology leadership, Micron’s memory solutions enable the world’s most innovative computing, consumer, enterprise storage, networking, mobile, embedded and automotive applications.

Advertisement

Follow Us:

Sitemap | Privacy | Copyright © © 2017, WSTA®, All Rights Reserved.