IBM and Cisco target ‘early 2030s’ for quantum networks

• Partnership will combine IBM’s expertise in quantum computers and Cisco’s in quantum networking
• IBM will convert stationary quantum information in a quantum processing unit (QPU)  into ‘flying’ quantum information 
• The aim is a federated architecture capable of trillions of quantum operations

IBM and Cisco have teamed up on a mission to enable practical, distributed and networked quantum computing “by the early 2030s”. 

Parlaying the individual skills and expertise of the two companies – IBM’s in quantum machines and Cisco’s in quantum networking technologies – the initial aim is, by the end of 2030, to demonstrate the first proof of concept (PoC) for a physically linked network combining hardware and software comprising “individual, large-scale fault-tolerant quantum computers” that “work together to run computations over tens to hundreds of thousands of qubits”.

Furthermore, the two will collaborate to “solve fundamental challenges towards a quantum computing internet”. If successful, that would be quite something.

According to Jay Gambetta, director of IBM Research (and IBM Fellow), the tech giant’s roadmap “includes plans to deliver large-scale, fault-tolerant quantum computers before the end of the decade. By working with Cisco to explore how to link multiple quantum computers… together into a distributed network, we will pursue how to further scale quantum’s computational power. And as we build the future of compute, our vision will push the frontiers of what quantum computers can do within a larger high-performance computing architecture.”

Vijoy Pandey, general manager and senior VP at Outshift by Cisco, the San Francisco-headquartered company’s incubation engine for emerging technologies, added: “Getting quantum computing to useful scale is not just about building bigger individual machines, it is also about connecting them together. IBM is building quantum computers with aggressive roadmaps for scale-up and we are bringing quantum networking that enables scale-out. Together, we are solving this as a complete system problem, including the hardware to connect quantum computers, the software to run computations across them, and the networking intelligence that makes them work.”

Quantum computers that are sited physically close to one another but which are separate devices can be linked together today under laboratory conditions, but scaling beyond that sort of level, which is being researched and tested today in various parts of the world, is problematic for a myriad of reasons.

Quantum computing is a very delicate process that is subject to many environmental and engineering vagaries and, in their collaboration, IBM and Cisco will endeavour to transmit qubits over longer and longer distances, such as between different buildings or datacentres. Such ability will be key to operating and managing networked distributed quantum computing. The two companies will also experiment with optical-photon and microwave-optical transducer technologies to determine whether they could and should be incorporated into a quantum network to transfer quantum information.

The plan for the initial PoC is to entangle qubits from multiple separate quantum computers located in distinct cryogenic environments. To do this, both IBM and Cisco will have to invent new connection technologies, including the aforementioned microwave-optical transducers and supporting software stack. It will be a major task and involve the definition of a complete hardware and software stack able to preserve fragile quantum states, distribute entanglement resources, facilitate teleportation between quantum computers, and synchronise operations with sub-nanosecond precision.

Multiple linked quantum computers will require a powerful and novel interface, and IBM says it plans to build a quantum networking unit (QNU) to serve as the interface to a quantum processing unit (QPU), with the explicit task of taking stationary quantum information in the QPU and converting it into “flying” quantum information through the QNU to then be further linked across potentially multiple quantum computers.

Meanwhile, and equally vitally, Cisco’s quantum network technology, if proven successful, will distribute quantum entanglements to arbitrary pairs of QNUs on an “on-demand basis” and so drive the quantum information transfer required for a given quantum algorithm or application. With that goal in mind, Cisco is developing a high-speed software protocol framework that can continuously and dynamically reconfigure network paths so entanglements could be distributed to the QNUs when partial computations have been completed.

New “network bridge” technology required

The companies will also work together to determine how a “network bridge” comprising new hardware and open-source software could use Cisco quantum network nodes to link many IBM QPUs within a datacentre through its QNU interface. In the future, this approach could be extended to link QPUs across multiple datacentres and scale a bigger quantum network across even longer distances.

It is hoped that such an ability would, in due course, constitute the foundation of a quantum computing internet. IBM quantum computers linked by such an architecture could enable massively computationally demanding workloads, including those that require high-performance computing resources as part of a quantum-centric supercomputing framework. Such a network would permit problems to be solved via potentially trillions of quantum gates.

Indeed, IBM is also working with the Superconducting Quantum Materials and Systems Centre (SQMS), led by the Fermi National Accelerator Laboratory, one of five research centres funded by the US Department of Energy as part of a national initiative to develop and deploy the world’s most powerful quantum computers and sensors. 

The SQMS comprises a cohort of more than 550 experts from 36 partner institutions, including from national laboratories, academia and industry. It is a “mission-driven, multidisciplinary collaboration that integrates deep expertise in quantum information science, material science, applied and theoretical superconductivity, computational science, particle and condensed matter physics, cryogenics, microwave devices and controls engineering and industry applications.”

Great aspirations, so much still to do

The hope is that a quantum computing internet will emerge where distributed quantum-based technologies, such as quantum computers, quantum sensors, and quantum communications, are connected and share information across distances, such as across a city, a region, a country, a continent and eventually on a global scale, to provide ultra-secure communications.

It’s a fine and laudable ambition but, as of now, the requisite hardware is little more than partially developed, the proposed new networking stack is so immature as to be little more than a gleam in a computer architect’s eye, and an enormous amount of further research into quantum physics is needed if quantum states across distance are to be stable and robust enough to carry information and do something with it. 

That said, IBM and Cisco firmly believe the best way to solve what is a massive problem is to treat it as “a system-level engineering effort, one that integrates hardware, software and network intelligence into a unified framework”. 

Their latest collaboration is a sign that quantum computing and networking are quickly moving forward to a stage where developments will no longer be measured in terms of increasing qubit counts but from the ability to connect quantum devices into a potentially huge network that is strong and flexible enough to solve the problems that mankind might not know even exist.

– Martyn Warwick, Editor in Chief, TelecomTV