#3 - Five-Nanometer Silicon Oxide Switches

Okay - I chose this new technology for two reasons. First, my inner geek just fell in love with the concept of utilizing nanotechnology for advancing material science and manufacturing, and second, because I believe that manufacturing innovations don't get as much credit as end user centric innovations such as software. With that said, this posting will discuss the expansion of semiconductor data bandwidth as a result of advancements in silicon oxide switching (transistors).

I assume that most of us within the IT industry are familiar with Moore's law. If not, Moore's law states that ‘the number of transistors that can be placed on an integrated circuit doubles every 24 months.” That means chips get more complex, faster, and smaller every two years. This law has been true within the semiconductor industry for the past 40 years and up to now, has hit a roadblock. The reason for this roadblock is that there have not been any recent innovations in semiconductor manufacturing to allow the addition of more and smaller transistors within a single processor. You might not have noticed the roadblock, but if you look at processors in computers today, you’ll notice that the trend for companies like Intel is to stack processors with multiple cores (processors within processors).  The reason for this is that manufacturing technologies cannot create a processor core smaller than today’s core, so semiconductor companies are now stacking (in a three-dimensional stack) their smallest processors as a workaround. 

However, since this blog is about innovation, there is an invention out of Rice University that not only allows processors to be created smaller, but will actually improve the fastest production processor data bandwidth (memory capacity) from 32GB to 1TB. This completely goes against Moore’s law as this is more than a doubling of transistors per processor; it’s an increase of 3,025%!!!!  Jun Yao is the credited inventor of the tiny five nanometer wide silicon oxide switch, which is conceptualized in the picture below.

The Silicon Oxide Chip Jun Yao/Rice University

Jun claims his invention came about by looking where the semiconductor industry was not. In fact, the solution was in front of everyone’s eyes. For the past 20 years, the industry has been focusing on reducing the physical strands of graphite transistors. Jun noticed that processors are already made of silicone, which also conducts electricity well, ….. so why not make a switch out of silicon?  (One up for thinking outside the box!) In any case, the trials proved successful beyond Jun’s expectations and now the industry is flooding Rice University and Jun’s team with grant money to continue their research. The current chip that is being tested has already exceeded 1TB of data bandwidth. 

So what does this mean for us users?  I don’t know. One thing that comes to mind is the impact of the computer manufacturing industry in the late 1990’s; there was a small wave of technology improvements in manufacturing that created more powerful processors, servers, etc. This explosion of powerful infrastructure components allowed now popular technologies such as virtualization to take hold and revolutionize the IT industry. Because an invention such as Jun’s is at the front of any technology wave, the impact cannot really be predicted and only truly understood by looking back years later. One can only speculate, but perhaps with more powerful processors, data centers will become even smaller, and cloud computing will become cheaper and more feasible with smaller and MUCH more powerful processing capabilities. Thin clients anyone?