This year’s installment of the Supercomputing conference series was held in downtown Dallas, a stone’s throw away from Pioneer Plaza with its bronze statues of cowboys and cattle. As if emulating the early settlers’ optimism, the conference oozed belief in scientific progress and technology to the extent that a group of declared techno-optimists added ribbons to their conference badges that read “Glass Half-Full”. Established in 1988, Supercomputing has become one of the most important dates in the calendar of high-performance computing experts worldwide in the past 30 years. Marking this anniversary by taking selfies with an exhibited Cray-1 computer, conference attendees were acutely aware of how far high-performance computing had come in the past decades and eager to discuss what the next 30 years may bring.
In this atmosphere, a panel discussion about “Quantum Communication Networks and Technologies”, organized by the Alliance for Quantum Technology’s Intelligent Quantum Networks and Technologies program, was bound to produce unusual and invaluable insights. The panels’ goal was to discuss the opportunities that quantum networks could afford and how the high-performance computing and quantum physics communities could collaborate to implement this technology. The panelists were:
- Mercedes Gimeno-Segovia, Lead Quantum Architect at quantum computing start-up PsiQuantum in the Bay Area
- Joseph Lykken, Deputy Director of Research at Fermilab and leader of the Quantum Science Initiative there
- Christoph Simon, Professor at the University of Calgary specializing in long-distance quantum communication among other topics
- Maria Spiropulu, Shang-Yi Ch’en Professor of Physics at Caltech and founder of Alliance for Quantum Technologies
- Jake Taylor, Assistant Director for Quantum Information Science at the White House office for Science and Technology, professor at University at Maryland and Fellow at the National Institute for Standards and Technology.
The discussion was moderated by Leonie Mueck, Division Editor for Physical Sciences and Engineering at PLOS ONE.
Like classical computer networks, a quantum network consists of quantum nodes, which could be quantum processors or quantum memories, connected by quantum channels. Single-photon technology at optical or near-infrared wavelengths allows for quantum states to be transmitted through these channels. Quantum networks are governed by the laws of quantum mechanics and, as a consequence, opportunities and applications differ profoundly from what classical computer networks offer. For example, entanglement can be shared between the nodes and used as a resource in ultra-secure communication or in enhancing the accuracy of measurements beyond classical limits by reducing noise.
“We don’t really know yet which opportunities may arise from quantum networks, but we see glimmers,” said Jake Taylor mentioning quantum-enhanced metrology as well as fully homomorphic encryption as promising applications. The latter would imply that a quantum computation in the cloud could be performed with perfect encryption of data and computation. When asked about the near-term “killer application” of quantum networks, Christoph Simon named quantum key distribution, a method of ultra-secure communication based on quantum states transmitted through optical fiber networks or free-space links. “Quantum Key Distribution is a mature technology with experiments running in Calgary and elsewhere,” he said, identifying China where an optical-fiber link between Beijing and Shanghai as well as a satellite for quantum communication have been launched as a major player. Simon voiced disappointment that the US is not investing more in this technology.
Mercedes Gimeno-Segovia highlighted distributed quantum computing as an important application of quantum networks. Much like in the classical case, a quantum computation could be spread to many networked quantum computers coordinating the different computation steps between them. Since qubits, the basic unit of quantum information, are very sensitive to noise, quantum computers will need to employ error-correcting codes by redundantly encoding quantum information in many qubits. One fully error-corrected logical qubits will consist of 100,000 to 1 million physical qubits, explained Gimeno-Segovia, and this sort of size could be reached with a distributed quantum computing architecture. “Large quantum computers will have to incorporate quantum networks in their guts: There will be small quantum computers acting as processing units which will perform distributed computing between them,” she remarked.
Bringing quantum networks from the drawing board to employment will require a well-functioning ecosystem of innovation and collaboration. “National Labs like Fermilab are at the forefront of creating such an ecosystem,” Joseph Lykken stated. He explained how the Fermilab’s expertise in producing superconducting cavities for particle physics experiments is being transferred to making cavities for storing quantum information in quantum networks. Maria Spiropulu, who is also a particle physicist by background, stressed that the collaboration between particle physicists and quantum computing experts in the context of National Laboratories has already led to tangible results: Together with quantum physicist Daniel Lidar, her research group has run calculations on a quantum annealer that extracted the Higgs boson signal from noisy data. “There is some skepticism but also a lot of enthusiasm about this collaboration in both communities,” Spiropulu said. Spiropulu also discussed the FQNET project, a scalable user facility for fundamental research on distribution of quantum entanglement and a testbed for technology developments and quantum device benchmarking for quantum communications.
When asked about the most important recommendation for an action that needs to be taken to make quantum network technologies thrive, the panel was remarkably united: bringing people with the necessary expertise to engineer, produce and scale these technologies together. This would inevitably include members of the high-performance computing community with expertise in classical distributed computing. While the audience was pleased to hear about these opportunities for collaboration, there was also criticism as to why their expertise was not represented on the panel. Taylor conceded that nascent efforts to bring these communities together in the context of societies such as the ACM (Association for Computing Machinery) need more attention.
Efforts to establish closer ties will certainly not fail due to lack of enthusiasm and optimism from the high-performance computing community. As a start, the audience did not consider the thought that quantum-mechanics based computing and networking concepts might soon turn the world of high-performance computing inside-out as futuristic fantasy but as a realistic opportunity: about a quarter of audience members raised their hands when asked whether they believe they will see a fully-fledged quantum internet in their lifetime.
Featured Image Credit:Fabrizio Carbone/EPFL