True Custom Computers are Back

Ask anyone in the 1990s or today about a custom computer and most will say the same thing. You take a motherboard, an Intel or AMD processor, memory chips, hard drive, power supply, optical drive and you snap them onto the motherboard and you have a custom computer. You can get custom computer kits from various suppliers in order to create a homebuilt computer for personal, home use or gaming. This build process proceeds at a larger scale such as when companies like Synnex builds custom computers for organizations in the same league as the likes of Facebook, Google, and An article on looks at this phenomenon of big cloud companies building their own custom computers rather than ordering them from DELL, HP, or IBM.

Further reasons for the pursuit of custom built computers is explored in an article on Ars Technica that discusses Facebook designing their own servers and having those computers made to spec at a massive scale. We do not look at servers here per se, but the wider movement, it seems, back towards highly custom computers rather than off the shelf models. We think aloud through some of the ramifications of this before we take a look at what the future of standardized computers may become.

Assembled Computers

A homebuilt computer is often described as one you personally create in a process that far more resembles combining Lego Blocks more than it does an engineering process. Making a custom computer is more about connecting premade parts together and not about creating those parts. Describing these machines as Custom-built PCs do not appear to be the right way to refer to these kinds of computers. A better term would be assembled computers.

Custom Computers Was The Way

Between the 1950s and 1970s, most or all of a computer was made from scratch. Every part of a computer was unique and there was not a lot of mass produced computers. You could not easily take one part out of a computer made by one company and replace a similar part in the computer made by a different company. The parts of a computer were unique to a given model and manufacturer. The secret designs of the parts were known only to the engineers that produced them. These were true, custom computers.

The Long Run of the Generics

What happened to the true custom computer? It was too expensive, to hard to make, and took too much effort for tech support. By the time the 1980s rolled around, the generic computer took over and true custom computers became less common. The computers most of us used that we today call desktops and laptops grew out of those generic standards from the 1980s. Generic computers led to a boom in nearly all things connected with computer technologies and computer networks.

Custom Computers Rise Again . . . Maybe to Stay

Now, it seems we are on the verge of going back to the age of true custom computers. Tablets and smartphones are computers but many of the parts that define them are not the standardized parts you can buy for laptops and desktops. Yes, tablets and smartphones have processors, data drives, memory, screens, and a main board to connect all these things, but each of those components are uniquely designed for each tablet or phone model. Unlike laptops and desktops in which you can swap out these parts with better replacements from third parties, you are generally stuck with the designed parts of a tablet or smartphone just like the custom computers of the 1950s – 1970s.

Uniqueness Creates Isolation

Even the operating systems and software is unique. Apple iOS devices are best programmed using tools from Apple that produce software that is not directly understood by tablets running Android or Windows. Software for Android designed according to the preferred guidelines and languages for Android will not run well, out of the gate on iOS or Windows. With the acquisition of Nokia, Lumia tablets running the RT version of Windows will have software incompatible with Android and iOS.

Compare all that to laptops in which you could routinely write software for Apple MacBooks, Windows, and Linux since those systems generally ran on standarized processors from AMD and Intel. Even then, it is challenging due to the differences among those operating systems, but it was an approachable challenge since toolkits that masked many of these differences could rely on the fact of a common hardware standards that, in turn, made those operating systems somewhat similar in general structure. Whereas the principles of iOS, Android, and Windows RT are quite distinct from each other owing to their ability to redefine hardware standards and certification requirements closer to their needs rather than in consideration of interoperability. This phenomenon of uniquely defined hardware and the supporting systems that goes along with it may be known as vertical integration.

What Others Say About Vertical Integration?

As stated earlier, companies once existed who made custom computer chips to run computerized systems. Then, Intel made the microprocessor available to the general market in which software could be written in such a way that custom, unique computer chips were less necessary. This caused an explosion in technology. This story is related by Richard Rumelt in chapter 13 of his book, Good Strategy, Bad Strategy. He shows how the introduction of the microprocessor as a general device for replacing custom hardware with flexible software allowed many new companies to spring up and reorder the technology landscape.

The transition from highly customized and tightly connected electronic components was the move from vertical integration to horizontal integration. The timeline of this transition is described in a Harvard Business Review review article by the name of Ubundling the Corporation where they look at the evolution of vertical integration among different types of organizations.

Given the challenges of vertical integration, why would anyone pursue it? It may be born out of a reaction based on a perception of how the market is evolving. Well, there are benefits to the company that pursue this course. Vertical integration can be a means to control the supply chain in order to better ensure the availability of parts and capabilities for a given firm. This method of vertical integration is described in a piece by the International Association of Business and Management Professionals. They show examples of companies such as DELL who have used vertical integration in this manner with varying degrees of tight control.

The long-term business challenges of vertical integration is well known. Analysts at the The Wharton School, University of Pennsylvania present a view that vertical integration may work for some but not all. You need a particularly well defined process to succeed with vertical integration. They contend that many who undertake vertical integration may eventually return to a horizontally integrated model when the returns on an end-to-end approach decline.

Vertical integration is a subject with considerable research supporting and discouraging its adoption. Such is the extensive review of the promise and peril of vertical integration from McKinsey & Company. A company can see cost and market control benefits with vertical integration but conditions in the market can eventually undermine future gains. The effort required to manage the operational parameters that support vertical integration can work against an organization. One of the leading causes of lowered gains from vertical integration has to do with shifts among customers, suppliers, and competitors. Shifts that are dynamic and unpredictable in which advantages may be conferred to those that are specialized.

Some See the Next Billion Needing a Tablet or Phone

The return to the true custom computer means that a single company can link together all the elements of an electronic device into a single product to maximize what they gain from its sale. Such is the implication of a report in which a senior leader at Microsoft describes the new preference for custom handheld computers over standardized machines. While this approach may be great for profit maximization, it may hold undesirable effects as far as the incentive to supply standardized components at a level that sustains standardized computers.

Vertical Integration as a Temporary Advantage

While highly custom computers made of parts and designs unique to single companies may provide faster, sleeker devices for a while, they eventually hit a dead end. As far as the companies that fully embrace full vertical integration, they may encounter a limited ability to act quickly to market changes and evolve technology at such a rate and scale of knowledge distribution that benefits the overall advancement of technology. This seems to be the contention set forth in an article from the Wall Street Journal that looks at the return of vertical integration. Several companies are cited in the article who were or are on opposite sides of the vertical integration question. The caution of adopting such an approach is briefly described in the article and it is worth considering when evaluating technology choices between those that are vertically integrated or best of breed.

The Apple Phenomenon of the 2000s

Basically, the Apple iPad started the trend towards the custom computer. My question is, why did we all follow Apple? Looking back, what seemed to have happened was a sense of fear developed in some parts of the computer industry that Apple would take over. As Apple’s fortunes rose and eclipsed the growth trajectory of other established players, there was a sense that the most effective response was to mirror the vertical integration philosophy successfully employed by Apple. I, was certainly a huge proponent of this and many things Apple. I still like what they have done but I do not see their particular outlook as the type of group think around which we should all rally.

Identifying Today’s Custom Computers

Despite appearances, Apple MacBooks are still mostly standardized computers with premium components. I single those computers out because it is important to understand what things really are versus what is marketed. Apple MacBooks along with all other laptops and desktops running Intel or AMD processors consists of parts that are interchangeable. You can take most of these machines and replace the parts in them with better versions. A slower hard drive or memory module may be replaced with a different one with better features and speed. On a desktop, a physics card could be added that allows new scientific discoveries to emerge out of an adhoc lab in a garage. New sound and graphics components could be added that allows artists to make masterpieces that are more stunning and informative than the more basic versions would allow. People who invest in an economy model computing device can stretch their investment with upgrades from a wide list of parts suppliers who compete on quality, cost, service and more.

The ability to take these machines much higher than were they started gives a person the opportunity to achieve newer and better results than the initial investment in the machines would offer. You can add or link parts that can provide even more capabilities and have new technologies available on multiple machines regardless of the operating system they run in the case of desktops and many laptops. This can save costs in some cases over having to struggle against insufficient systems parameters to reach goals with greater quality and depth. The significant upgrade-ability, redefinition, and modification of hardware allows the owners of the machines to extend the value of an investment considerably. As such, the nature of this standardization and tremendous flexibility has set conditions that has help enable intellectual exploration and technology proliferation to expand in a substantial way over the last 30 years.

Cloud and Mobile Emphasized

That all seems less implausible with the march to extensive vertical integration centered on closed, unique, highly custom computers. Intellectually and commercially, I can appreciate what Microsoft calls a mobile-first and cloud-first direction. It is a massive shift in how computing is done and offers a scale of flexibility that many of us can appreciate. What you loose in that shift however may be far more valuable than what you receive in exchange. What mobile and cloud solves are traditionally difficult problems in computing in terms of seamless cross machine availability and system resource capacity. Full local, on premise systems like desktops, laptops, and the traditional software that supports them do provide many benefits but it can be hard to reconcile the benefits of local systems with those of the cloud/mobile and not find yourself sliding towards the latter. Microsoft tried this for a while with a strategy called Software + Services but it was not a competitive strategy in terms of execution. It is easier to do either all on premise or all cloud/mobile.

The trio of Microsoft, Apple, and Google seem to all be headed in the same direction. All three have cloud strategies with varying levels of emphasis, but they are largely the same. All three use the cloud as background for their Apps offerings. Most of what you can do through Google’s Android presence, you can do purely through the Web. Some of what you can do through Microsoft’s you can do on the Web. A very small amount of what you can do through Apple’s App presence, you can do through the Web. Apple places the most emphasis on apps, followed by Microsoft, then Google. Each have operating systems and each have custom computers for those systems with Apple having singular control over their custom computer, while Microsoft has a flagship offering but with other companies able to use their operating system. Google sees the widest range of devices with their operating system, but they are all custom nonetheless.

Regardless of their differences, both Google and Microsoft’s offerings are generally modeled after Apple’s iPad. Among the 3 of them is the same strategy with small differences. A custom computer as a gateway to indefinite subscription based services on the Web. A recurring revenue model for computing environments.

The model is realized differently by each organization using similar custom mobile platforms and cloud mechanisms. Apple largely gains recurring revenue through a share of the sales of content and apps delivered over the cloud to iOS devices. Microsoft achieves the same primarily through access to cloud infrastructure for IT and first party apps for productivity work on Windows. Google makes it work through information resale in which first party apps and the Android environment can be a gateway to individual level telemetry of various kinds.

Eventually, Microsoft will expand to do more with individual level telemetry and content while Apple may expand into infrastructure and productivity and before it is over, the only difference between them being the logos, fonts, color schemes in the operating environments.

I really like Apple technology and I agree with most of their design and engineering results. Yet, as I step back and look at a global level, I see a trend where you have no way to judge if the way they do things in the future will be the right things or the right choices if other choices are absent. What if a future Apple computer used a type of hard drive that is fast for video editing but it could have been faster, more secure, and high capacity at the same time but I would never know it because competitor hard drives no longer exist because only 1 company makes custom computers and they decide what hard drives to use? The economic incentives they face simply means that what they offer is no longer relevant for the computer engineers, scientists, and technology theoreticians. Artists may have an adequate machine, but not one that situates them in such a way as to push the envelope of creativity because the technology is too closed and inaccessible to standardized expansion.

Between the Apple iPad, the Nokia Lumia and Android tablets, we have useful set of custom computers that work well for what they are designed to do. However, while these computers are very convenient, we are loosing functionality if we choose them exclusively in place of desktops and laptops. It is highly unclear that we will be able to afford future standardized computers should these custom hand held computers become more dominant. That may be overstating things however, and such an outcome is yet to be seen on the horizon.

The Life Computer

Steve Jobs was right to assume that a content device was what most truly wanted or needed. A standard computer was designed to do more than what it was routinely used to do. To surf the Web. Ironically, for a time, it was the only device that everyone could buy in which people who consumed content and those that made things with it could have access to it through common distribution channels. Since it was the only design, the acquisition of the devices by the large majority of those that used them for information reception and communication essentially subsidized the continued availability of the devices for the small number of persons who used them to productively create, explore and analyze.

These over powered machines while applied towards passive and casual activity was an opportunity in waiting. Should an individual get an idea suddenly in the middle of the night, they could turn to these otherwise passive machines to then do what they were really designed to do and make an idea a reality. It was always an opportunity in waiting, but one that simply did not see such opportunities engaged. Now, we are seeing the uptake of computers that we would admittedly recommend as a Web, communications, light content tweak and full entertainment engagement device.

With such a single purpose, perhaps custom fitting the design of the computer towards more limited functions was an appropriate course to take. At least these things such as entertainment, communications, and Web use are the optimal set of functions to focus on rather than general purpose expandability and cross compatible standards support.

Co-Existence of Custom and Standardized Computers

The major problem for the continued availability of standardized computers is that as more people adopt custom computers in the form of smartphones and tablets, the manufacturers of hardware will decide not to make them. What could happen instead is that a new generation of standardized hardware may arise. Hardware such as Arduino, Raspberry Pi, and Intel NUC could be harvested to allow people who need them to create custom desktops, laptops, and servers based off these micro boards. Newer and smaller organizations could emerge who do not have the profit mandates of today’s large technology incumbents to provide systems that allow creation, discovery, and engineering activities at a level that matches or exceeds what current standardized hardware enables.

As I look at the landscape of hardware options involving Raspberry Pi, Intel NUC and Arduino, alongside custom, specialized computers in the form of today’s tablets and smartphones, I see a repeat of history. I am reminded yet again that history does have a tendency to repeat as if people are fully captured into a cycle of doing what was done before although it looks different. In the 1960s and 1970s, companies customized and narrowed their designs in such a way to increase performance and make it difficult for competitors to match the design’s functions. At the time, the small development that was the microprocessor was hardly on the radar, but it grew out of seemingly nothing and changed everything. It was not the only time, that happened, I am also reminded that throughout the last 100 years, similar waves in computing has taken place.

In that light, many of us may lament the loss of the standardized computers with which we had become familiar. Yet, the indications derived from such initiatives as Raspberry Pi, Arduino, and Intel NUC shows great promise and perhaps far greater flexibility. Just the other day, someone took one of these highly general computing technologies and created a cell phone. That is certainly far more than what you’d see in laptop technology and maybe the shift towards unique, proprietary, mass distributed custom computers in the form of smartphones and tablets means that more space opens up for the new wave of future standardized systems that holds perhaps more promise for technological development than their predecessors would have encouraged. That is an encouraging thought indeed.

By Michael Gautier


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s