The New Open Opportunity

India is emerging as one of the world’s biggest electronics markets. For instance, it recently overtook the US as the world’s second biggest market for mobile phones, behind only China. While this is in general a good thing, the downside is that it is projected to lead to a huge trade deficit. India is expected to consume $400 billion worth of electronics by 2025, but will only produce $100 billion worth. Therefore, there is a thrust aimed at creating electronic products in India itself.

One element of this is the creation of many incentives by the Union government’s Department of Electronics and Information Technology (DeitY) to encourage the production of electronic products in India, for India. The aim of the “Startup India, Stand Up India” programme is to create entrepreneurial ventures and DeitY would like to support those that focus on electronics. There have already been several efforts underway: electronic incubators, of which DeitY has funded three. In addition, it is envisaged that there will be another 35 incubators and at least some of them will support electronics companies.

All this is happening at a time when things are changing dramatically in the electronics industry. I wish to make an analogy between what happened in software due to the advent of open source and what is happening in the hardware industry today. In both cases, the dominant paradigm that was proprietary and owned by large companies has been turned on its head by the advent of new open source paradigms.

The sea change in software happened some time ago, when there was a transition from the proprietary world as exemplified by IBM and Microsoft to the open world exemplified by Unix and later Linux. In effect, what had been a closed and insular world became far more democratic. The metaphor of “cathedral and bazaar” captures this idea rather well. The net result was an explosion in creativity and an upending of the old order. It was a truly disruptive change, in the sense used by Clayton Christensen, an innovation guru.

Earlier, software product development had been something that could only be done by an army of engineers with the huge resources of a large company at their disposal. This was because you had to write the code specifically to a hardware platform, and you needed to delve deeply into the intricacies of the hardware. You had a “stack” that went all the way from the peculiarities of the CPU (central processing unit) architecture to the topmost layers of the software, and this essentially had to be rewritten from scratch for each new machine that came out. That, in fact, was the “soul of a new machine”, as in the Pulitzer prize winning book by Tracy Kidder.

This changed when standard operating systems were developed and provided to third party software developers, as in the IBM 370 OS, DEC PDP 11 and VAX OS. DEC or Digital Equipment Corporation, although now a footnote, was in the forefront of this change, and it freed software developers from worrying too much about the intricacies of hardware. They could depend on a Hardware Abstraction Layer (HAL) provided by the machine makers, and expect a consistent set of functions to be provided to them.

But still, the operating system was written by hundreds of engineers and was essentially impossible for any one person to completely understand, because it was monolithic and had obscure legacy code and hidden assumptions and interlinkages.

The truly revolutionary changes were twofold: one, the invention of a well-architected and modular, portable operating system, Unix, written by just two legendary engineers, Ken Thompson and Dennis Ritchie, at Bell Labs. The second was the opening up of the software code, first of Unix, and later its derivative Linux (which is Unix rewritten with no code owned by Bell Labs, along with the GNU code and its copyleft, sort of the opposite of copyright ©).

Microsoft went part of the way down this track, making its Windows portable, but it is still monolithic and certainly not open. The two reasons are why in the new world of devices, Windows is nowhere, and Android and Apple’s iOS, both derived from Unix and Linux, now dominate. It appears as though, in keeping with the Second Law of Thermodynamics, entropy increases and the “bazaar” dominates.

Today, almost no software developer needs to worry about the hardware. They can use freely available stacks, some of which are fully open source, such as LAMP (Linux, Apache, mySQL, php). This frees them to worry about the problem to be solved for the customer, and not the infrastructure. Thus it is easy to write new applications.

In my opinion, the world of consumer electronics is experiencing a similar, and largely positive, shift. The traditional process of creating a hardware platform is roughly as follows: you have an idea for what your electronic circuit should accomplish. You first use eCAD tools to create a portable description of the circuit (in some ways reminiscent of the intermediate representation of the semantics of a piece of code by a compiler) known as Gerber files. This representation can then be fed into machines to create the printed circuit board and then to populate it with the components you have designed in. At the end of the process, you have a finished board, which you then encase in a mechanical enclosure for shipment.

In passing, let us note that provision of a platform is a powerful incentive to improve the use of the underlying technology. Just as the hardware abstraction layer in the software world helped hide the complexity, there is a similar mechanism developing in the world of block chains. At the annual Netexplo innovation conference in Paris on February 10, I was glad to see an Israeli startup named Colu win a prize for creating a block chain platform, on top of which all sorts of applications can be developed, not just bitcoin.

The quasi HAL being developed for open source hardware is similarly a powerful mechanism that relieves the application provider of the burden of understanding the details of the underlying hardware. Instead, they can feel free to focus on the customer’s problem. In effect, if the electronics developer stays with a standard hardware environment like Arduino or Raspberry Pi, it is primarily a programming task to create the application. You can be assured of support for the standard interfaces regardless of the specific implementation: it could be a button-sized version meant for a wearable device, or a full-sized board.

This has two interesting implications for electronics innovators.

First, if you have a good idea, you can put a fully functional prototype together relatively easily, and go to market or at least field test it. Time to market, always so crucial for a startup, can thus be reduced.

Second, you can then go for an optimization activity and redesign your electronics from scratch. The obvious reasons would be to reduce costs and also generate intellectual property. By creating your own electronics, you can avoid competing with others using standard boards, which may have many features you are not using.

This redesign can be a significant business model for Indian service providers good at frugal engineering. That is, they can offer product design optimization services which could include both electronic and mechanical design. The value added in the product often is in design. And “Make in india” is not only about manufacturing but also about designing. Much like software services firms are able to benefit from the outsourcing trend, this could be a business model for design houses.

Thus, the changing environment in electronic system design and manufacturing (ESDM) opens the door both for innovative new companies, and for those who want to offer design services.

Rajeev Srinivasan has worked for innovative companies such as Bell Labs, Siemens and Sun Microsystems in strategy and product management. He has taught innovation at several IIMs, and is a member of the Think Tank working on India’s national IP policy. He has been a conservative columnist for almost 20 years, and has degrees from IIT Madras and Stanford Business School.