“Quantum computers have gained the powerful abstractions that allow programmers of classical computers to design and integrate new apps and hardware, and connect devices into networks with ease.
Abstractions are the keystone of the dynamic ecosystem of information and communication technology. You can think of them as allowing you to drive a car without understanding how the engine or transmission system works. For most drivers, a car is just a box with a steering wheel, pedals and a gearstick. The details of whether it is a full-electric vehicle or a model with a conventional engine, say, are largely irrelevant.
Read the paper: An operating system for executing applications on quantum network nodes
In computing, abstractions in the design of operating systems and interfaces mean that a company or research group does not need to reconfigure transistors in a computer’s processor, or design an entirely new way to interface with the Internet, to make an innovative idea happen. Instead, developers can just plug their innovation into pre-existing, standardized schemes.
But what’s true for conventional, classical computers has not, up until now, applied to potentially one of the most disruptive innovations in information technology — quantum computers. That changes with the work of Delle Donne et al.1. The authors introduce a set of abstractions and standardized interfaces for running quantum applications distributed across devices in a quantum network. This is a major advance on previous test-bed implementations of quantum networks2, which required custom hardware and software tailored for each specific set-up. As of now, the wheel does not have to be reinvented every time.
‘Quantum internet’ demonstration in cities is most advanced yet
To understand a little more about the importance of abstractions in computing, let’s take a step back. The end user of a classical computer — be it a smartphone, a laptop, a data-centre server or a game console — will generally run applications such as messaging or word-processing apps, or web servers or games. But these applications don’t interact directly with the hardware. Instead, they rely on an operating system to provide a simplified, uniform interface to resources such as the central processing unit, working memory and peripheral gadgets such as hard drives or webcams.
The idea of process abstraction is central to the design of any operating system (Fig. 1a). The operating system does not care about the internal details of the applications it is running, or what they are doing: they are just ‘black box’ processes that interface to it in a uniform way. Similarly, each application runs without ‘awareness’ of any other application, as if it has exclusive control over the hardware. Process abstraction is how modern operating systems enable multitasking, running multiple applications as simultaneous processes without them interfering with one another.
Even the operating system itself does not interact directly with hardware, but does so through specialized software modules called drivers. These drivers represent another abstraction: they offer a standardized interface that allows operating-system developers to support various hardware technologies — whether an old magnetic hard disk or a modern solid-state one — without rewriting core software, simply by adding a new driver module. This is the principle of ‘implementation isolation’, which emerged in the 1960s as mainframe computers, which were originally designed for single tasks, began evolving into modern multitasking machines. The transformation was driven by foundational concepts introduced by Dennis Ritchie and Ken Thompson at Bell Labs with their UNIX operating system, which continue to shape computing today3.
The AI–quantum computing mash-up: will it revolutionize science?
Applications that need to communicate over a network use another key abstraction: network sockets. Just as an electrical socket connects devices to the power grid, so network sockets allow applications to send and receive data across the Internet, independent of the underlying network transmission technology, whether it involve fibre optics, copper cables or wireless links.
Quantum computing and quantum communications have recently made significant progress, raising expectations that an Internet of networked quantum devices might be imminent (or perhaps not). But quantum technologies operate under fundamentally different principles from those that gave shape to modern computing systems and network protocols. Unlike classical bits, quantum bits (qubits) cannot be copied or shared freely, owing to a feature of quantum mechanics known as the no-cloning theorem. Moreover, remote interactions between quantum computers depend on entanglement. This is a phenomenon with no direct classical counterpart, in which operations on one qubit instantaneously affect another, even at a distance4.
These differences necessitate new abstractions that are still as intuitive and flexible as those developed for classical computing. The QNodeOS operating system developed by Delle Donne and colleagues1 is a theoretical and experimental framework that provides these for networked quantum computers (Fig. 1b). In QNodeOS, an application consists of intertwined quantum and classical procedures, or blocks. This architecture enables developers to use quantum computing resources only when needed, while offloading routine computations to classical systems.
The quantum internet has arrived (and it hasn’t)
Within a quantum block, developers can perform local operations on a quantum processing unit, or establish entanglement with remote qubits through a quantum network processing unit (QNPU). Much like the Internet Protocol stack, which regulates access to the Internet in classical networking, the QNPU abstracts the complexity of quantum communication by offering developers quantum sockets, with a uniform set of operations independent from the underlying network technologies. Furthermore, the principle of implementation isolation is maintained: the QNPU does not access quantum network hardware directly, but instead interacts through quantum devices. Each of these has its own driver — itself abstracted — allowing for distinct underlying technologies.
To validate their approach, the authors demonstrated a simple, yet representative, network application: delegated quantum computation, in which a task is split between two quantum computers. This involved generating a pair of entangled qubits spread across two quantum network nodes, a precursor step to quantum teleportation — transferring a qubit state between two nodes by destroying it at one and recreating it at the other. Entanglement generation was made possible by introducing quantum sockets, an abstraction enabling seamless quantum communication. QNodeOS can also manage concurrent execution of multiple quantum applications on the same quantum computer through a process scheduler. The hardware independence of these abstractions was demonstrated by implementing QNodeOS on two different quantum processors: one using nitrogen-vacancy centres in diamonds5, the other a trapped-ion system based on a single calcium atom6.
Nobody can say for certain whether QNodeOS will become ‘the new UNIX’. The success of abstractions and interfaces depends largely on how well they align with the needs of mainstream applications — something that is highly uncertain at this early stage of quantum computing. However, QNodeOS is a step in the right direction and can lead quantum computing and networking into a new phase of maturity. By establishing standardized abstractions, researchers and engineers can accelerate progress, replicating the transformative impact that classical computing and networking abstractions had on technology, industry and society.” [1]
1. Nature 639, 312-313 (2025) By Claudio Cicconetti
Komentarų nėra:
Rašyti komentarą