Moore’s law turns 50: And why it still counts?

Guruswamy Ganesh

In April this year, one of the most popular phenomenon in the semiconductor industry turned 50 –The Moore’s Law. Calling it a phenomenon is not wrong, given it is the most quoted and misquoted phrase to explain and predict the way forward for the chip industry. It all started with the observation by Gordon E. Moore on the number of transistors in an integrated circuit doubling approximately every eighteen months.

The law has had several iterations as the timeline for doubling the transistor count kept changing. From 1965 to 1975, the transistor count on a chip doubled every year. In 1975, it went to two years and back in the late 90s, the storage industry accelerated it again. However, since 2008, the pace slowed down leading to doubts being voiced about the possibility of this law remaining relevant in the future.

What makes the Moore’s law work is an unusual combination of factors. Shrinking transistors improves performance by reducing signal distances. More high-performing small chips can be plucked out of a wafer of a finite size as against the earlier larger, slower counterparts, which means that the same fixed capital costs can generate more revenue.

Moore’s Law drives the innovation we take for granted every day – from smartphones to the Internet to big data. Even after 50 years, it is amazing how this “Law” continues to ring true for CPU and memory performance, allowing them to continue to scale to meet the needs of today’s data hungry environments. It has been used by the industry to govern roadmaps for form factors and performance for years.

People at large, enjoy the benefits and convenience of devices shrinking with better performance year after year, without realizing the significance of Moore’s law in action in the background. Semiconductors are about as cool to the outside world as vacuum tubes they replaced. If not for guidance such as this, an iPhone would have been the size of a steamer trunk, cellular relay stations would rival the Qutab Minar in size and a Google datacenter would consume as much energy as Mumbai city. And that is the beauty of this concept which is the epitome of simplicity and complexity coming together.

The biggest and most underappreciated achievement of Moore’s Law is that it institutionalized optimism in the industry. It brought predictability to the world of technology highlighting that despite the high costs and numerous challenges, technology would invariably be a better, safer investment than virtually almost anything else. The fact remains that investing billions in new fabs, or energy efficient datacenters, entails risk, but far less than drilling for oil in the Arctic. The law also brought healthy competition taking the semiconductor industry forward because if any player slowed down, the other would take over.

Let’s take a look at Moore’s law in the context of the storage industry. Over the years an exponential gap has grown between the advancements of CPU and memory, and what traditional storage can handle.

As CPU, memory and networking all continued to follow Moore’s Law and double performance about every two years, hard drive performance is the laggard with Moore’s Law helping hard drive density but not performance. In fact, in some cases, performance has actually slowed down as capacities grew. It was more than 10 years ago that the first 15K RPM hard drive was introduced into the market, but we have yet to see 30K or even 20K RPM drives. The reasons for that are the mechanical limitations that prevent hard drives from achieving faster rotational speeds. Information on a hard drive is read from and written by a combination of mechanical parts, the actions for which may take time – and therefore introduce delays.

There are several strategies deployed to counter mechanical delays. Server and Storage vendors invest heavily in controllers that use processors, DRAM and other techniques to work around the hard drive bottleneck, but those can only help so much. Over time, this difference in performance creates major inefficiencies, requiring re-architecting or rewriting applications for balanced system utilization. But with flash, the fix is easier, more reliable and more cost efficient!

Solid-State Drives (SSDs) have brought storage back in line with Moore’s Law. For instance, just four short years ago the average SSD achieved around 250MB/s throughput and capacity topped out at about 512GB. Today, enterprise-grade PCIe application accelerators such as those from SanDisk achieve 2.7GB/s data transfer, and reaching up to 6.4TB of capacity in a form factor that fits into the palm of your hand.

Though the initial costs of implementation of SSDs are higher than HDDs, they offer cost savings in the long run for businesses with their lower energy usage and greater productivity with higher Input/outputs Operations per Second (IOPS). One SSD delivers the performance of 100 hard drives. When very little power is needed to operate SSDs this in turn leads to significantly less heat output is generated by the systems.

So, what does the future look like? Researchers have said that we may hit a fundamental physical wall with regard to semiconductor shrinkage by 2021. However, 3D memory will continue the virtuous cycle. Instead of squeezing transistors tighter together, you stack them up and gain (if not increase) the economic/technical benefits. SanDisk’s 3D BiCS memory contains 48 layers, a high water mark for the industry. For consumers and large businesses alike, smart devices and data center equipment will become more intelligent because far more memory can be squeezed into Smaller footprint.

(The author is the vice president, SanDisk India)