Open source hardware and software are the new version of a phenomenon that’s been with us since the first wandering bards added their own words to the melody of a well-loved folk tune. Open source tools feed the dreams of amateurs and entrepreneurs alike from a deep well of well-documented designs, which are constantly being refined by the communities that use them.
A growing number of traditional businesses and corporations understand that embracing them as an integral part of their business strategy may be their best hope for surviving the disruption and disintermediation which open source technologies often bring to tightly-held markets.
At first glance, the idea of freely-shared intellectual property (IP) would appear to contradict the fundamental principles of a competitive, market-driven economy. Open standards that enable interoperability between competing products are acknowledged as tools for building larger markets, but tightly-held IP and proprietary knowledge have been considered the key ingredients for innovation and success in technology based markets from the steam age onward (see box: ‘Open Standards vs Open Source’ for details).
This has begun to change over the past two or three decades, as the growing complexity of many technologies has given rise to several trends that threatened to slow the pace of innovation and market evolution:
- Many products are built upon layers of basic hardware and software functionality that are so complex it would be impractical or impossible for most companies to develop themselves.
- Likewise, the resources needed to detect and fix the bugs in those technologies, or update them to meet evolving market requirements, are beyond the resources of all but the largest single company.
- Innovation is further discouraged by corporations that routinely create or acquire IP with the sole purpose of creating barriers to market entry for potential competitors.
These issues have helped drive up the cost of creating or licensing many essential IP elements to the point where they make many ventures less lucrative — or even impractical. Paradoxically, the stifling business conditions they created also served as the stimulus for the emergence and adoption of open, royalty-free technologies that are developed and are supported by the people who use them.
This emergence occurred in three distinct phases, beginning with open source software. The open hardware movement which followed somewhat later, adapted many of the concepts and licensing conventions developed over a decade by the software pioneers.
Recently, open source silicon platforms have begun to emerge as the third phase of the movement that is re-democratising innovation and powering the global tech economy.
A tale of two innovations
Phase 1: Linux
One of open source software’s biggest successes is the Linux operating system, conceived by Linus Torvalds as a free alternative to AT&T’s powerful but costly UNIX OS. Fittingly, Linux borrows many elements from Richard Stallman’s earlier GNU project. Within a few years of its initial release in 1991, Linux had attracted a large, active community of users, many of whom began contributing to its vast and growing pool of user-created resources.
As a result, Linux developers can access countless libraries of pre-written utilities and functional blocks, created by fellow users. With many of the low-level elements of their applications already available, developers are able to focus their efforts on the proprietary functions that add value to their product and differentiate it from its competitors. Linux’s ability to shorten a product’s time to market while reducing development costs made it a top choice for the embedded systems that run ‘under the hood’ of thousands of consumer and industrial products. In addition, Linux powers the servers that run 96.5 per cent of the top one million domains in the world (as ranked by Alexa), 92 per cent of Amazon’s 350,000 EC2 Web servers, and all 500 of the world’s most powerful supercomputers (https://www.makeuseof.com/tag/linux-market-share/).
Phase 2: Arduino
The Arduino embedded computing platform was at the vanguard of the second open source wave (hardware) when it was first conceived in 2003 as a student project at the Interaction Design Institute Ivrea in Ivrea, Italy. The project’s audacious goal was to develop a very low cost, easy to program, single-board computer that would enable designers, artists and other non-engineers to “…build digital devices and interactive objects that can sense and control objects in the physical and digital world” (https://www.arduino.cc/en/Guide/Introduction). In retrospect, the developers significantly underestimated the project’s actual impact.
The original Arduino board was roughly the size of a business card and provided easy access to the signal lines of its 8-bit microcontroller through a series of pins along each side of the board. These pins allow a developer to easily connect the processor’s input and output lines to a real-world device, such as a light, sensor, or motor. They can also be used to connect drop-in daughter boards (known as shields) equipped with sensors, network interfaces, or other functionality. Programming the boards is accomplished using simple, English-like commands that are written and compiled within a user friendly integrated development environment (IDE).
Arduino’s low cost, versatility, and ease of use led to its explosive popularity with the technology hobbyists who embraced the do-it-yourself (DIY) philosophy of the so-called Maker movement. Sold in the millions, the tiny open source board and its variants served as the silicon intelligence that enabled hobbyists to unleash their imaginations and create things like low-cost automated telescopes, smart clothing, personal robots, and automated home brewing systems.
Meanwhile, as enterprising makers began to offer innovative commercial products on Kickstarter and other crowdfunding sites, Arduino technology helped them bring their ideas to market quickly and inexpensively.
A powerful legacy: As some of the earliest open source platforms, Linux and Arduino were instrumental in upending the conventional top-down business model in which proprietary technologies were developed in isolation by large corporations and doled out to those who could afford them. In contrast, opensource’s bottom-up approach enables designers to harness the collective intelligence of a user community to create tools and products that reflect real-world needs.
The overwhelming success of Linux and Arduino also demonstrated that product differentiation on a common, non-proprietary platform is not only possible, but desirable.
Open source technologies perform another vital and surprising role within the tech economy by serving as a catalyst that enables the creation of other open source technologies that make new and unexpected applications possible. In much the same way that Stallman’s GNU project provided many of the concepts and enabling technologies that made Linux possible, Arduino gave makers, tinkerers, and would-be entrepreneurs the low-cost, easily hackable controller they’d need to tackle projects that might have seemed unthinkably ambitious until then.
Among other things, the humble controller board provided the spark that ignited the 3D printing revolution. Although the first 3D printers were developed in the 1980s, they were still costly and relatively rare, typically found only in large corporate R&D labs. This changed in 2005 when a team at the University of Bath in UK began a project to develop a low-cost 3D printer that could print most of its own components. They named the project RepRap, short for replicating rapid prototyper. The machine would melt inexpensive plastic filament and extrude it through a print head that would be moved across the X, Y and Z axes using inexpensive stepper motors. Each of these motors was controlled by its own Arduino board, while other Arduinos took care of various housekeeping functions (https://www.digikey.com/en/articles/techzone/2011/nov/arduino-open-source-platform-unleashes-creativity).
An early RepRap prototype printed its first useable part around a year later, paving the way for a subsequent design that was able to print more than half of its functional parts in February 2008. Tech enthusiasts from around the world quickly embraced the RepRap project, using the open source design to build their own machines, and contribute improvements that they’d made. Soon, entrepreneurs were using RepRap’s ‘genetic code’ as the basis for affordable commercial products, such as the Thing-o-Matic, a sub-US$ 1000 3D printer kit, produced by MakerBot. The Thing-o-Matic, and virtually all of the other early machines, used Arduino controllers, and most of the low cost units sold today use Arduino-style controllers.
The Arduino also played a seminal role in the development of low-cost multi-rotor and fixed-wing drones that enjoy wild popularity with both hobbyists and commercial operators. Much of the early activity of this revolution occurred, and is documented on the DIY Drones website (https://diydrones.com), an online forum for hobbyists interested in developing autonomous model aircraft.
Phase 3: Open source silicon
Although open source technologies gave developers royalty-free access to hardware and software platforms that they could modify to suit their needs, the integrated circuits their platforms run on have been almost exclusively based on proprietary IP. This began to change as devices known as field-programmable gate arrays (FPGAs) became inexpensive enough to be used as the equivalent of a custom integrated circuit. Designers can program the devices with the desired functionality using large libraries of FPGA code (both proprietary and open source) in a matter of days or weeks, as opposed to the months and millions of dollars required to develop the tooling for a real IC.
Although FPGAs are excellent solutions for certain applications, they are slower, more power hungry and significantly more expensive than an equivalent single-purpose silicon device. Until recently, nearly all ‘hard’ silicon designs were based on proprietary IP, either developed in-house or licensed from a third party vendor. The cost and time required to develop a reliable, high-performance processor architecture meant that nearly all ICs that require a processor core use third-party IP, licensed from a handful of vendors such as ARC, ARM, MIPS and Qualcomm. These cores offer excellent technical support, optional architectural enhancements, and deep libraries of application code, but the IP tends to be very costly to license and usually has significant restrictions on how it can be modified.
As a result, the same market pressures that created open source software and hardware are giving rise to the first generation of open source silicon projects. Among the most ambitious and successful of these efforts is the RISC-V project, an open instruction set architecture (ISA), originally developed at the University of California at Berkeley, and currently supported by the RISC-V Foundation (https://riscv.org/).
Like most modern ISAs, it is based on reduced instruction set computing (RISC) principles that provide a good balance between silicon footprint, execution speed and code efficiency. The instruction set can be instantiated with a 32-, 64-, or 128-bit word width and has been designed to efficiently support a wide variety of common applications. Unlike proprietary ISAs, the RISC-V ISA may be freely used for any purpose, enabling anyone to design, produce or sell RISC-V chips and software.
The Foundation’s 100+ member organisations have pooled their knowledge, know-how and resources to create a common platform and an ecosystem of development tools that form the basis of their own unique and highly competitive products. SiFive (https://www.sifive.com/), for example, is developing a service that allows its customers to design custom RISC-V based System on a Chip (SoC) devices and receive sample chips within a matter of weeks. SiFive also enables newcomers to become familiar with the technology by offering a low-cost, Arduino-compatible board powered by one of its RISC-V processors.
RISC-V’s customisable architecture makes it easy for developers to create devices with the precise mix of functionality, performance and energy-saving features that give their products the unique capabilities they need to dominate a particular application space. As an example, Trinamic, a German semiconductor maker focused on motor drivers, is adopting the RISC-V as the foundation for its next generation of motor drivers and power electronics. The amount of computational power required for motor control functions has been steadily increasing to meet the faster response times, higher precision, and greater overall performance required by many emerging applications. The RISC-V architecture makes it easy to add cores to the driver ICs that are responsible for the communication and protocols, offering more complete solutions and higher levels of integration.
The road ahead
As the RISC-V and other early open source silicon platforms emerge, their development ecosystems will get a head start from the rich trove of resources already created by the open software and hardware movements. By enabling small, agile groups of entrepreneurs to bring cutting-edge products to market quickly and in a cost-effective manner, these open source technologies will continue to democratise and accelerate innovation within the tech economy.
