- US - English
- China - 简体中文
- India - English
- Japan - 日本語
- Malaysia - English
- Singapore - English
- Taiwan – 繁體中文
Last month, I was on a plane to Asia to meet with customers for our quarterly technical reviews (QTRs). During these QTRs, we align technology trends, product strategies and innovation projects we’ve all been working on. Somewhere over the Pacific Ocean, I was groggily reviewing SSD performance trends and couldn’t help but wonder what the need for these ever-increasing product specifications was. For context, the SSD performance (sequential reads) has been doubling almost every two to three years. Currently, we have consumer drives on the market that saturate the PCIe® Gen5 bus, topping off at a blistering 14 GB/s of sequential read performance.
I was trying to figure out which existing PC applications or use cases demand such high performance. The simple answer was — not much right now. At first glance, it looks like we’re overengineering our SSDs for applications that don’t yet exist widely. The overwhelming majority of PC users won’t come close to consistently realizing these SSD performance metrics in their everyday use cases.
How ‘Dune: Part Two’ helped me figure out SSD performance trends
As I let this thought swirl in my head, I closed my laptop. Unable to sleep, I decided to rewatch one of my favorite movies for perhaps the fourth time, “Dune: Part Two,” Denis Villeneuve’s directorial masterpiece.
As the visually stunning scenes unfolded, I couldn’t help but admire the director’s grand vision in bringing Frank Herbert’s book to life. The book came out in 1965 and was an instant classic. There have been several attempts to capture this epic saga on the big screen, but they’ve all been lackluster. One primary reason has been that the technology of the time never allowed for such complex thematic elements to be visualized elegantly — until now.
That’s when it occurred to me: If not for the unrelenting march of technology powered by our collective innovations across multiple domains, this movie could never have been made! This thought forced me to reconsider my original question. Instead of asking which PC applications currently exist that demand such high performance from our SSDs, we should be asking “What future applications could benefit from such performance?” When we think of the issue this way, we encourage a more optimistic view of the future versus a cynical view of the present.
This naturally led me to my next thought — dancing honeybees. Hold on while I explain!
What dancing honeybees can teach us about the need for innovation
Click image to enlarge
I recently learnt about a rather fascinating behavior honeybees exhibit — the waggle dance! When a solitary bee discovers a new source of nectar, it “beelines ” back to the hive and does this remarkable dance to tell its colony mates the precise direction and distance to the nectar source.
However, scientists observed another interesting behavior, which was reported in a recent study in the Science journal. Around 80% of the bees followed the instructions and headed directly to the nectar source. However, the other 20% did something very counterintuitive. Rather than following the instructions, they followed paths not communicated in the original dance. This seemingly aberrant behavior allowed these maverick bees to discover new nectar sources instead. It makes sense that this behavior has significant evolutionary benefits in ensuring the colony's survival in case their primary food source is ever depleted or destroyed. Colonies that fostered this exploration thrived while the rest inevitably perished.
The lesson we can learn from the humble honeybee is this: Exploration of the unknown and innovation aren’t merely nice to have; they’re must-haves. Fortunately, Micron is one of those places where innovation is ardently fostered in many ways. The company has been issued over 57,000 U.S. patents and counting, and innovation is an integral part of the culture here.
There has always been a cyclical relationship between hardware and software: Significant strides have been made in the hardware realm with no apparent use cases, but over time, software that can use the hardware emerge and soon the pace of software advancements outpaces the current hardware capabilities. Then innovation is needed to catch up. The throughline has always been looking beyond what’s currently required or even possible.
This has been our mantra when it comes to designing our SSDs. In a highly commoditized segment like client SSDs, it’s relatively easy to stick to a customer-defined specification and build identical products like everyone else. But it takes a lot of courage and importantly a culture that values innovation to build differentiated products. Recently, Prasad Alluri, general manager of Micron’s client SSD business, wrote about how we’re championing storage innovations for the AI PC.
Besides these advances, there’s a roster of ideas we’re constantly pathfinding, prototyping and experimenting with, the results of which I will certainly write about. Some of these ideas might very well never see the light of day, but there are bound to be some that lead us to yet undiscovered nectar sources.
