The Copernican Theory Of Computing

The compute horsepower in your hands is nothing compared to what’s available across the network.


The most powerful computer on the planet isn’t sitting in some research center being cooled by liquid nitrogen. It’s sitting in your pocket or on your desktop—or at least a piece of it is.

We’ve all been brainwashed into thinking that it’s important to own the most powerful computing device you can get your hands on. Latency, after all, is annoying. Somewhere over the past couple of decades we have been transformed into a society of speed freaks whereby existence is measured in milliseconds. A decade ago we were comfortable watching pages download from a 56kbps modem. Now we want an entire movie instantly.

What’s interesting is that we can now achieve this kind of I/O intensive computing on even the smallest device, but we still tend to think of it from the inside out. It’s not just the device that’s making this better. It’s also not the clock frequency of the CPU inside it. Most CPUs are powerful enough to run just about any application fast enough to satisfy most consumers. After all, a smartphone is a powerful computer in its own right. But it’s only a small piece of a global network that reaches well beyond just a single device or even a single company. It’s not even a single cloud. It’s a conglomeration of clouds—more like clouds upon clouds.

This is the new supercomputer. It has an almost unlimited number of cores, incredible bandwidth, a huge number of security holes and some of the best power management capabilities available. It’s a grid or mesh network, with peer-to-peer capabilities rolled in. What the individual user controls is their own access, their portion of the power consumption, and the amount of money they contribute to making this giant compute farm work.

Consider a social network such as Facebook, for example. While a sizeable portion of the core database runs Facebook’s servers, even more is processed inside the devices used to load pictures and communicate with friends. If Facebook had to process all of this information it would grind to a halt. But through a network of computers, whether they’re classic PCs and laptops, smartphones and tablets, all of this work gets spread out and done in small bits—and extremely efficiently. Breaking down the silos of information also breaks down the silos of computing.

History may never actually repeat itself, but there are some disturbing similarities that crop up over time. One is viewing the world the way it has always been viewed, which in Copernicus’ time was from the Earth out rather than using the Sun as the starting point. We’re at a similar juncture in computing, which could have a big effect on how much compute power we will need in the future, where it should reside and what we can do with all of this capability when we need it.