Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

The rise of domain-specific machines: Rethinking the future of computing

As technology advances and security concerns grow, the need to rethink how we design and implement computing systems has become urgent.

The rise of domain-specific machines: Rethinking the future of computing

[Adobe Stock / N7]

Profile photo of Majeed Javdani

President · International Digitalization Forum

As technology advances and security concerns grow, the need to rethink how we design and implement computing systems has become urgent. The evolution of programming models and hardware architectures points to the need for a return to domain-specific systems that align closely with business logic, security needs, and industry standards.

FROM BARE METAL TO HIGH-LEVEL LANGUAGES

Before the emergence of high-level programming languages, machines were embedded with hardware-specific enabling languages designed for a specific logic. This contrasts with the abstraction principle of high-level languages, which allow programmers to implement various business logics in the same machine.

In the early information processing era, machines operated with basic operating systems or none at all, requiring programmers to directly manipulate hardware using low-level instructions. The runtime environment at that time was primarily bare metal, meaning programs ran directly on the hardware without an operating system.

The implementation of a programming language results in a virtual machine, which should enable a software programmer’s creativity to build their business logic rather than focusing on how the underlying hardware machine works.

The driving principle of high-level languages was to provide programmers with a level of abstraction, allowing them to implement their business logic using human-readable syntax, while leaving the translation into machine-readable syntax to supporting tools like optimizers. What optimizers do is optimize the algorithms that constitute a business logic, aligning them with machine-specific instructions.

THE PROMISE—AND PITFALLS—OF MULTIPURPOSE MACHINES

Multipurpose machines emerged with the rise of PCs in the 1980s. The personal computing phenomenon successfully achieved its mission of building an information ecosystem, but the downside of this advancement was the proliferation of personal computer usage in business.

Before the proliferation of the internet and connected machines, the risks associated with information security were minimal, and using PCs in businesses was an efficient approach to information management. However, by connecting these multipurpose machines to the internet, businesses exposed themselves to numerous threats.

IT’S TIME FOR A NEW MODEL

Considering the advancements in hardware production, the rise of advanced processing chips, including the emergence of open-source processing instructions like RISC-V, and the proliferation of AI, now is the right time to consider a new evolution in the design and implementation of programming languages.

These advancements may prompt us to reconsider an old information processing model, bringing it back to life in the new age of computer science. The revival of domain-specific machines presents an appropriate solution to the heightening concerns of business leaders regarding information security risks.

ALIGNING INFRASTRUCTURE WITH BUSINESS LOGIC

Moving from the utilization of multipurpose machines to deploying domain-specific ones is a critical risk control that should be implemented in business transactions.

The data structure of purpose-built machines is defined by the business logic instead of abstract data structures. The data control and storage management structures of these machines follow the data structure of the business logic, which represents a specific industry or transaction type.

The deployment of heterogeneous systems across different industries and business sectors could decrease the likelihood of widespread cyber attacks, and it is not contrary to alignment with standards.

DESIGNING FOCUSED ENVIRONMENTS

From a technical perspective, the implementation of domain-specific programming languages should result in a runtime environment that acts as a virtual machine directly invoking machine instructions. On the other hand, it should provide a focused IDE to the programmer for implementing transaction-related algorithms.

When I say “focused,” I mean the data structure in these implementations is purpose-built. The programmer has only the tools necessary to model the algorithms required for business logic in alignment with the target industry.

Processing instructions are nothing but primitive data types and supporting operations specific to each processor family. In the context of purpose-built machines, runtime libraries will leverage these hardware-specific data types and operations to dynamically structure domain-specific data types.

THE ROLE OF ENTREPRENEURIAL INNOVATION

The rise in market share for chips based on open-source instructions like RISC-V depends on the development and adoption of operating systems, compilers, and runtime libraries that are compatible with these instructions.

Small tech firms should build a business ecosystem for products based on open-source processing instructions, from the chip itself to purpose-built firmwares, file systems, runtime environments, compilers, and focused IDEs.

When we talk about a purpose-built machine, its DNA will be inherited from the characteristics of the target business domain and industry.

By mobilizing small, agile teams, tech entrepreneurs can utilize programming language standards and implement industry-aligned, domain-specific development tools and runtime environments.

STANDARDS-BASED IMPLEMENTATION

What is critical for entrepreneurs to understand here is that standards like ISO or ANSI are open, non-proprietary instructions. They are copyrighted documents, but the instructions are free for utilization and implementation.

For example, the C programming language has ISO and ANSI standards and has already been implemented as different proprietary tools, such as Visual C++, C++Builder, and others. Standards are available to everyone, can be leveraged in alignment with the creativity of the implementer, and they are free—but your own implementation becomes a proprietary asset.

This way, entrepreneurs ensure the conformance of their products to the standards, as well as interoperability with other standard-based products. Meanwhile, their product will constitute their future assets, which are capable of valuation and various forms of monetary utilization.

Following industry standards enables an ecosystem with less effort. A software product that is implemented using adopted standards seamlessly provides an environment where other standard-based products can interact with it—which means interoperability. Interoperability can then lead to the evolution of your product through what others create.

Standard-based implementation is a risk control for market entry, as business executives and decision-makers are often reluctant to buy nonstandard products. The procurement of standard-based products is a risk control that businesses can implement to reduce risks associated with the integration, support, and maintenance of software products.

A STRATEGIC SHIFT FOR INDUSTRY INNOVATORS

To meet the challenges of tomorrow, business leaders should embrace purpose-built computing aligned with open standards and domain-specific logic. This shift isn’t just a technical upgrade—it’s a strategic move for innovators ready to shape the next era of secure, efficient, and industry-driven software ecosystems.