When we attended Cloud Expo 2012 in New York City, we took the opportunity to conduct a survey on trending IT topics, gathering responses from a variety of vendors and other conference attendees.
We found some survey results to be quite interesting. To our question “Do your most critical business applications respond as quickly as you’d like?” 54% said “No.” Of those attendees who said “No,” 61% reported that their company is not using solid-state devices (SSDs).
It was not entirely surprising to find that 54% of Expo attendees are not achieving the performance they expect, because it’s an often-heard complaint. But even more interesting was the result that, of those experiencing poor application response, 39% were using solid-state devices.
So, why are some companies that deploy solid-state technology experiencing performance issues? Isn’t the point of Flash to attain optimal performance for critical applications?
The simple answer to these questions is that not all SSD is created equal.
Solid-state storage offers the promise of high performance for a wide range of enterprise applications. Unfortunately, some SSD vendors have caused confusion with bold claims and positioning of their solutions that have not always lived up to advertised expectations during actual use.
SSD technology is available in several variations, including Direct Attached Storage (DAS), hybrid arrays and all-Flash arrays. To realize the expected performance of SSDs, IT professionals need to consider the requirements, attributes and features needed to support their data center. The key is to not only select the type of NAND Flash memory used in the drive, but also to evaluate the intelligence of the built-in SSD controller. The controller is the command center for NAND memory, assigning the location where each memory cell will read or write data and for communicating with the interface that connects to the computer – it is a critical component to ensure the expected performance, cost, efficiency and reliability of an SSD.
NAND Flash memory types include single-level cell (SLC), multi-level cell (MLC) and enterprise multi-level cell (eMLC). Depending on circumstances, any type of Flash can have endurance or reliability challenges, particularly for write-intensive and mission-critical applications (e.g., online transaction processing, social networking, e-commerce, cloud computing) because Flash memory wears out with repeated writes to individual cells over time, unless the correct tools and technologies are implemented for optimizing SSD endurance and functionality.
SSD endurance is the usable life of the Flash memory cells (i.e., the number of writes a cell can maintain). Every electrical charge passed through a NAND Flash memory cell as part of a read or write operation wears down that cell. Because Flash memory cells will be written to many times each day, endurance determines the reliable life of each drive.
Endurance, performance, reliability and availability of SSDs depend directly on the design of the SSD controller and not on NAND Flash memory, as many believe. How effectively this controller manages the Flash memory determines whether the SSD can truly handle enterprise applications that require continuous uninterrupted operations with heavy read and write workloads.
An advanced controller architecture that analyzes and adapts dynamically to the changing characteristics of the Flash cells as they age can improve the performance and extend the life of NAND Flash memory. SSDs that rely on older controller technology, such as HDD controllers, are not suitable for the continuous duty cycles and heavy workloads required for today’s demanding network environments, a potential reason for that unhappy 39% of our survey responders.
Ensure that your enterprise works with vendors who populate NAND-based SSDs with technology that enhances Flash and yields the desired capacity, performance level and cost.