A recent Wikibon article caught my attention called “Flash and Hyperscale – Changing Database and System Design Forever” because it echoes what I’ve been hearing lately about Flash Storage.
(For those of you not familiar with Wikibon, it is an online, professional community that features some insightful, moderated sessions and short research notes on challenging business technology problems.)
In exploring the economics of flash storage, the article addresses three objections often heard that show a profound misunderstanding of computing systems in general and storage systems in particular:
These are common doubts that I’ve heard expressed at client meetings and on tradeshow floors.
But what caught my attention are the conclusions drawn in the research article –that current applications and application designs are severely constrained by IO (we call that the big data bottleneck), and that low-latency IO reduces the constraints. They present graphs, research results, and case examples to demonstrate that businesses are becoming 20 percent more efficient with flash technology.
The author, David Floyer, believes this is a profound change in IT and the potential value IT can bring to organizations – I agree, based on the benefits we have seen from Ciber’s Big Data storage solution powered by IBM’s FlashSystem technology.
I also agree with his projection that this trend will start with mid-sized companies, who will move faster to ensure competitive advantage. I’m generally not so agreeable – but based on the first-hand results we’ve witnessed by implementing FlashSystem, it’s easy to support their conclusions.
Do you have other objections you’d add to the list above?