Business is Moving Faster than Moore's Law

Lori Mac VittieDistinguished Engineer and Chief Evangelist

Moore's Law is one of the most cited observations regarding technology of the 21st century. For those not familiar or in need of a refresher, Moore's Law refers to the observation that the number of transistors in an integrated circuit doubles about every two years. In layman's terms, Moore's Law is often expressed in terms of compute power doubling every two years.

A less technical perspective is that Moore's law is evident in the difference between the original [Star Wars or Star Trek] special effects and today's special effects. What used to take years to process now takes weeks or even just hours. It is why today's Oregon Trail would look very different than the pixelated version most of us grew up playing.

The importance of Moore's Law to business has always been in its impact on applications. The scale and speed of applications is tied to the compute power available. Doubling available power means more speed and greater capacity (scale). Both capabilities have been and remain critical to the ability of business to leverage its application capital.

In the past, Moore's Law constrained business growth to the pace of technological change. Coping mechanisms such as scaling strategies have helped business overcome the limitations but could not eliminate it. There is only so much you can do with the compute available, after all.

If you're old enough, you'll remember the CPU wars - when Intel and AMD spent a great deal of money to convince us their CPU was faster and better than the other ones. New CPUs were announced to the world with the same fanfare and giddy anticipation as that of today's iThings.

Today, only the most dedicated fans of hardware are aware when advances are made in computing power. There are no ads, no jingles, no songs. No fanfare. That's because new architectures and operating models mean business can move faster than Moore's Law.

By harnessing the power of containers and cloud and the reliability and speed of modern networks, we don't have to wait two years for compute power to double to increase our ability to process data and make decisions faster. We can double - or triple - computing power right now by distributing processing across as many nodes as we need. We can actually use machine learning in real time to provide business with the analysis and insights it needs to make decisions today instead of tomorrow.

That kind of capability is unprecedented. In the past, analysis took hours if not days. Today the same analysis can be arrived at in seconds. We use machine learning today for real-time threat analytics, user-behavior analysis, and behavioral DDoS detection. We use it to perform facial recognition and take advantage of biometric security. We use it in ways that would not be possible if we remained beholden to Moore's Law.

And that's a good thing, because physicists have already predicted the imminent collapse of Moore's Law. Theoretical physicist Michio Kaku says that by 2022, Moore's Law will be obsolete. We'll have reached the end of our ability to increase transistor densities and, with it, our ability to double compute. But we'll still have modern architectures, algorithms, and distributed systems that can double, triple, or quadruple compute in an instant.

The drawback is that Moore's Law enabled us to seamlessly double our compute. Rarely have we been forced to change anything - except for that switch from 32 to 64-bit systems - to realize the performance and capacity gains. That's not true for the alternative, systems-based approach that requires a hands-on approach. Sometimes that means refactoring applications. Other times it means new models of deployment and scale. It may mean rewriting systems to take advantage of parallelization or distribution. Regardless, it generally means change and effort to realize better scale and speed.

We don't have much choice now that business is moving much faster than Moore's Law can support anyway. That's why it's important that we continue to reimagine everything - from security to application delivery to operations. We have to move from relying on Moore's Law to increase our capacity and speed to relying on a system of systems that scales itself to process more data, more frequently, and faster than ever before.

Because we need to do more than enable business to make decisions in real time. We need to be able to anticipate the need to make those decisions in the first place.

About the Author

Lori Mac VittieDistinguished Engineer and Chief Evangelist

More blogs by Lori Mac Vittie