Greg Schulz (@ storageio) posted today about a UCSD study on the future of SSD/NAND Flash disk. As with anything Greg puts out, it’s a well thought out, sourced, and informative post. I’ve come to really value Greg’s opinion, and there is a very good reason he has the reputation within the industry he has today. *cough* Rockstar *cough* While I’m on the subject of pimping out Greg’s awesomeness, do yourself a favor and go buy his book. Cloud and Virtual Data Storage Networking, I picked up my copy at VMworld this year and it’s about as close to a tech bible as you can get.
So now onto my .02 cents.
Predicting the future in the technology space about the longevity of specific technology can be a fool’s errand, but that doesn’t stop marketers and academics from trying to do it.
For the record, I do not have a problem with academics sticking their fingers into current business and technology trends, my only quibble would be, many do not have any hands on experience on the implementation and management side to fully understand the “hows” and “whys” of a technologies impact. The saying goes, those who can do, those who can’t do teach, this while simplistic, does hold an inkling of truth. Many academics spend time in a cocoon of their own making; Cocoons in my view are not great incubators of ideas. Knowing a technology is all well and good, but unless you have felt the pressure involved in making that technology work in a production setting it may be difficult to fully understand the total impact of its use. Academics usually won’t lose their job if their assumptions or theories are wrong, technology implementers will. Getting a B on your paper won’t cost you your livelihood.
I speak for myself, but perhaps others in the implementation and evaluation side when I say that “markatecture” is a dirty word. The prevalence of the MBA class that has moved into technology from the marketing and message crafting side are able to fool the C-Suite, but those of us in the trenches can usually see through the slick packaging, the buzz words, and the hype. The marketers will take the research that the academics have done, and then utilize it as a club for which they will beat the competition and the customer over the head with. Email campaigns, mailers, and most evil the fake technology leader website. Much of the message is crafted towards the decision makers within organizations, and not so much the end user/implementers.
There is a very valid reason that most technology companies employ Sales Engineers as the tip of the spear towards those who are the day to day implementer’s of technology, and why the pure sales animal is the only person the CFO/CTO will ever meet during the sales cycle.
Predicting the future:
If I could tell you what the future held for “technology X” 2 years from now, with certainty, I wouldn’t be sitting here in my kitchen typing away on my laptop. The fact remains, that no one can prospect out that far in the technology space, especially when it comes to the Enterprise. Technological advances can outpace and displace entire companies. Case in point, anyone remember STEC? Yeah me neither. In less than 18 months they went from being the SSD tech darling, to an also ran and their stock took a nosedive reminiscent of Red Hat and Yahoo circa 2000. The very same fate could follow for Fusion-IO, though I think they are making good strides into diversifying their offerings. It’s a tough business, and it changes fast, and sometimes those changes are adopted, and sometimes they are not
Academics who want to try to foretell the death or demise of a specific technology, when they have zero actual hands on experience within the space would do well to temper their outrageous claims.
Same for marketers, we know you’re trying to spin “Product X” in the best light, but be mindful, the second you make a claim that is so patently false as to be laughable, your credibility is gone, and your sales will quickly follows.
If I had a dollar for every time I’ve been pitched the death of tape, I could retire. Remember SEPATON? Yeah how’s that working out? The death of the HDD is currently being bandied about, so I have an equally skeptical view of SSD’s having a “bleak future
Looking at reality:
Does someone really need a PhD to come to this conclusion?
The technology trends we have described put SSDs in an unusual position for a cutting-edge technology: SSDs will continue to improve by some metrics (notably density and cost per bit), but everything else about them is poised to get worse. This makes the future of SSDs cloudy: While the growing capacity of SSDs and high IOP rates will make them attractive in many applications, the reduction in performance that is necessary to increase capacity while keeping costs in check may make it difficult for SSDs to scale as a viable technology for some applications.
It more or less smacks of common sense and conventional wisdom, but because its come from an academic wing, it will be seized by the marketers to push a message.
Let’s face it, not every company is a Hedge Fund running HFT algos requiring dark fibre and sub 1ns latency, but many companies do have a need to address high IOP workloads, and for now, SSD’s and NAND Flash fit those needs well. If the cost per IOP fits within their ROI, then there is going to be a very high adoption rate at which SSD’s will be implemented. As the density vs. performance factor comes closer and closer to equilibrium, then SSD’s will continue to gain in market share and dominance.
SSD has its place. It’s a maturing technology that will probably be with us for the mid to long term. Will it replace Hard Drives? Yes, in some instances they already have, but not all of them. But just like Hard Drives have not replaced Tape, SSD’s will not replace Hard Drives, and to claim a “bleak future” based on a just exposes a short sighted view of an industry that both Marketers and Academics have a hard time grasping.