Network Computing is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Will Solid-State Storage Kill Tiering?

I had a comment brought to my attention the other day (it was from Twitter so it had to be true) that as the data center becomes increasingly populated with solid-state storage, technologies like tiering will become a thing of the past. In my opinion, nothing can be further from the truth. Solid-state storage will not only need tiering as it begins to be implemented in the data center, but also as it becomes the dominant type of storage in the data center.

First let's be clear, while in my last entry I predicted that solid-state storage may become the dominant form of storage in the data center sooner than expected, that conversion is not going to be complete tomorrow. It'll take the better part of the decade for solid-state storage to dominate. Even at that point, while it is the predominant form of storage, mechanical storage and, of course, tape will still have a major role to play in your storage infrastructure.

This time to convert is going to require bridging technologies that have the intelligence to move data to faster-tier storage as needed. While direct placement of data to the solid-state tier is ideal, in today's dynamic data center, staffed with stretched too thin IT administrators, the reality is that automation is the best bet in making sure that solid-state storage is used to its maximum efficiency.

After the conversion point, when over 50% of the primary data resides on solid-state storage, there will still be mechanical storage. In fact, it's reasonable to expect that from a sheer capacity standpoint that mechanical storage may be larger than the solid-state storage. Again, technologies like auto tiering and caching will be leveraged to move data to the mechanical tier.

Even when we reach the point where greater than 50% of primary data is on solid-state storage, or if we ever get to the point where there is no mechanical storage, there certainly will be tiers of solid-state. Today we have DRAM, SLC flash, eMLC flash, and MLC flash. Each of these variants of memory-based storage have their advantages. DRAM for example, while being the most expensive, has the best write performance and is the most durable. SLC flash is the most reliable flash memory and the MLC variants make solid-state storage more affordable.

Auto tiering and caching technologies are already being designed to take advantage of each of these different types of memory advantages while at the same time trying to avoid their weaknesses. There will be tiers of memory-based storage in the future just as there are tiers of hard drive based storage today.

The result of all this: Not only will solid-state storage not kill tiering and caching technologies, it will actually make them more commonplace. First, as the easy entry into the data center and then finally to leverage multiple types of memory-based storage within the data center. The pressure on vendors cannot simply stop at the caching or tiering technologies that we have today, but that pressure will force them to continue to enhance those technologies so that they can support multiple types of memory.

Follow Storage Switzerland on Twitter

George Crump is lead analyst of Storage Switzerland, an IT analyst firm focused on the storage and virtualization segments. Storage Switzerland's disclosure statement.