Keep Track of SSD Reliability

Arthur Cole
Slide Show

Changing the Way You Purchase Storage

Ensure that IT has the flexibility to build and efficiently run a shared infrastructure.

Now that a good portion of the enterprise community has embraced solid-state disks (SSDs) as key components of their storage infrastructures, it's time to ask an important question: Do you trust them?

The issue of trust is probably the most important overlooked aspect of SSD technology, given that its speed and power consumption advantages over traditional spinning media have helped propel Flash technology to some fairly important roles in the enterprise. And while nearly every technology lives up to expectations right out of the box, how long can you expect that trust factor - reliability in storage-speak - to last?

As Storage Switzerland's George Crump pointed out this week in an article on InformationWeek, SSD reliability can vary depending on the demands placed on them. High read/write environments will naturally lead to greater wear and tear, ultimately resulting in failure. As a general rule, however, single-level cell devices (SLCs) have the greatest longevity with around 100,000 read/write cycles. Multilevel cell (MLC) life spans are dramatically shorter - about 5,000 cycles - while the newer Enterprise MLC (eMLC) units can handle about 30,000 cycles. And beware of claims that "wear leveling" techniques will extend SSD life. They only serve to spread data evenly across all Flash cells, not to increase the number of cycles.

Drive manufacturers continue to utilize wear leveling and a host of other techniques to improve the reliability of their products. OCZ, for instance, touts everything from power loss protection and error correction to SMART monitoring and native command queuing as boons to reliability, while Intel has placed its stock in a hardware/software caching system called Smart Response Technology (SRT) for its new Larsen Creek model. There's certainly nothing wrong with any of these approaches, as long as we're clear that these are designed to improve reliability within the normal SSD lifecycle, not to extend the lifecycle itself.

Building storage arrays around Flash technology is also yielding new ways to boost reliability. EMC's Project Lightning, for instance, utilizes server-side and array-side Flash designs to help prepare drives for incoming data, much the same way processors signal each other in server configurations. The system is currently qualified only for Intel's PCIe Flash technology, however.

Solid-state storage is still a valuable addition to enterprise data environments, particularly when speed is of the essence. But it's always helpful when deploying a new technology to take full stock of its strengths and weaknesses. And in the case of SSDs, that means maintaining constant watch on their performance as time goes by.

Add Comment      Leave a comment on this blog post

Post a comment





(Maximum characters: 1200). You have 1200 characters left.




Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.