Microprocessors No Longer Valid Benchmark of Tech-Industry Success

There is nothing more disheartening than being an enthusiast of everything ‘technology’ or ‘innovation’ and hearing your novice coworker explain how "nothing is happening in tech" because of an article they read in Financial – insert name here – Magazine said so. While it is true that several CPU manufacturers are trading within 5% of their recent 52 week low, growth in the technology industry cannot be based solely on the financial health of chip manufacturers. In fact, most of the innovation in this space has deliberately abstracted itself from the reliance on computing hardware entirely. So why do we continue to place so much emphasis on the quarterly earnings of Qualcomm and Intel? 

It was in 1965 that Gordon E. Moore, co-founder of Intel, made his observation that the amount of transistors packed into an integrated circuit has doubled every two years. While the discovery was merely posed as an observation of historical data, it was eventually solidified as Moore’s Law. This law has continued to be relied on as a trusted resource in predicting growth and development. In recent years however, Moore’s Law has become completely irrelevant. Quantum computing has far surpassed the law’s time constraints, and as all computing hardware gets cheaper, the entire growth promise that Moore proposed in ’65 is dead.

At the outset it would appear this means the future is bleak, but as we peer out onto this beautiful technology landscape, the smell of virtualization and fresh blooming automation perks my ears. Okay, maybe it’s not quite that exciting. It does, however, suggest that the tide is changing, as it should. If you have read anything about the Internet of Things or the Internet of Everything, it is easy to imagine a world where every static/unintelligent device we use today becomes an interactive, connected, intelligent, and even predictive resource. This prospect should offer great promise in the world of processors right? Wrong.

Eliminating Unnecessary Processor Reliance

The Central Processing Unit is just that… it is central. It will always be at the center of our technology just as our brains will always be the center of our thought processing. More importantly, the idea that may change someday is far too scary for me to consider right now. What is not necessary any longer is having the processor present in every single endpoint. Instead we are seeing the processor, along with device storage, becoming centralized to a single control point. Much of the software, networking, and computational development has focused on efficiency, or in other words, using less computational hardware. So what is killing processor sales?

1. Centralization

Simply put, several inefficiencies are eliminated from a computing environment when the resources are centralized. One of the more obvious advantages is that it becomes easier to manage. I can have one technician repair the majority of my problems without ever leaving his location. I can also keep energy costs to a minimum. Whatever the advantage may be, the fact is that companies are doing it, and they are doing it rapidly. In the example of the Internet of Everything, the endpoints are often times just sensors that receive information and send it back to a central cluster somewhere that can actually make use of the data.

Some may argue that overall information security is reduced when the servers are centralized by giving hackers and other evildoers one throat to choke. It also should be mentioned that processing and storage hardware should be centralized to two or three locations instead of just one. By doing this you can ensure that no single failure can disable the entire infrastructure. 

2. Virtualization

The term virtualization is almost as vague as saying ‘the cloud’. There are several amazing advancements in virtual reality and network topology segmentation, but that is not what we are referring to here. Computer virtualization effectively encapsulates what you think of as your computer, or computing environment, and detaches it from the actual hardware. This means you can use your environment on one computer today, and another tomorrow. A huge database server may require thirty processors at 2 pm but only two at 7 pm. By implementing virtualization, companies can have much greater insight into what they are actually using and purchase hardware based on those metrics. In the past, organizations were forced to purchase for the worst case scenario. This allowed chip manufacturers to keep costs high and the grip on their clients nice and tight.

3. Software Defined Networking

Computers are not the only devices that use processors. Networking equipment accounts for a large portion of processor sales. After all, what is the value of a computer without internet or the ability to interconnect with other applications? In the past, networking technology has been ever reliant on, and constrained by, processor technology. SDN is changing that. This new development allows a network administrator to modularize each function of the network (routing, switching, stateful packet inspection, etc..) within a software suite. Now they can make changes to the network without having to be so reliant on the hardware.

Cannibalism

Cannibalism is real, it is scary, and it is more present today than ever before. We are seeing more often that the CPU manufacturers themselves are focusing their efforts on computational efficiency. Where at one point in time, new developments were all about more power (queue the music!), it eventually become about making better use of what you have (…). When the likes of Applied Materials and AMD begin focusing their efforts on how you can use less of their own product, you know the tide is changing. We also know that Innovation continues to surge and businesses are in-fact spending money on these technologies. With that said, we now know the money is not always sitting under the biggest roofs in Silicon Valley.