Last week, I came across an article in SearchSolidStateStorage that truly caught my eye. How couldn’t it with a title like, “When using SSD is a bad idea!” While there certainly were some valid points, I kept thinking the proper response was “never!” I wanted to use this post to respond to, and provide added context to, a few of the scenarios outlined in the article, as I believe that some recent advancements in SSD and Flash endurance enhancement technologies may have been overlooked when the article was developed.
Scenario #1: Don’t use SSD when applications are not read-intensive.
While at one point this was true, it is no longer an issue. By employing endurance enhancement technologies and by looking at Flash slightly differently, you can significantly improve write performance and Flash’s longevity. Take for instance our very own Guardian Technology Platform. With this intellectual property, we can significantly extend the endurance of MLC flash-based SSDs to deliver up to 50 full drive capacity‐writes/day for 5 years – a level that was once considered SLC-only territory. We can also tweak our SSDs to meet various endurance levels without changing the hardware, making them geared towards read- or write-intensive applications in the process.
Scenario #2: Don’t use general-purpose SSD in highly virtualized environments.
As with any storage decision, it is highly important to choose the right SSDs for your data center. Because of higher write activity in highly virtualized environments, SSDs geared for read-intensive applications would be a poor choice. They simply wear-out too quickly, and lead to a higher total cost of ownership. However, this does not mean that you have to use a more costly media (i.e. eMLC or SLC). Solving this problem is exactly what we set out to do two years ago. The Guardian Technology Platform allows us to drive additional endurance out of consumer-grade MLC Flash. By doing so, we can take Flash typically made for read-intensive applications, and make it relevant in the virtualized environments, reducing end users costs across the board.
Scenario #3: Don’t deploy consumer-grade SSD for enterprise applications.
I couldn’t agree more! First, let’s define a consumer-grade SSD as one that uses consumer-grade MLC Flash, is designed for consumer usage scenarios (i.e. typically read-intensive), and doesn’t include additional enterprise features. Deploying consumer-grade SSDs for enterprise applications is a surefire way to exhaust a SSD in record time. In enterprise server and storage applications, consumer-grade MLC Flash wears out quickly if left untreated. However, consumer-grade MLC Flash can be treated! As I mentioned earlier, our Guardian Technology Platform increases the endurance of MLC Flash so that it can reach enterprise endurance requirements. The Guardian Technology Platform also adds enterprise features like back-up power circuitry and enhanced error detection and correction, giving enterprises the data and drive reliability they necessary server and storage environments. Consumer-grade SSDs lack these critical features, making them even less suitable for enterprise applications.
Do you have any feedback on the points made in the article, or those I presented above?
In the end, I truly enjoyed reading the article as it brought up some good points that we in the solid-state storage industry are continually working to improve upon. One day, we hope to be able to say, “Using SSD is never a bad idea.”
Until next time.
- Bernie Rub, Vice President and Chief Technology Officer