Data Centers Need Design Flexibility for Quantum Computing Preparedness: 'Integration Capability Will Become Increasingly Important'
TECHWORLD ·
✦ AI Summary
As data center infrastructure design shifts to keep pace with the spread of AI, flexibility to accommodate quantum computing is also being required.
Gartner said design is needed with new computing platforms such as quantum computing in mind, not just AI.
Cisco and Fortinet mentioned post-quantum encryption, Cryptographic Agility, and the need to prepare for long-term operations in response to quantum computing threats.
As data center infrastructure design standards shift in line with the spread of AI, quantum computing is emerging as a new variable in the post-AI era. According to industry sources on the 31st, data center infrastructure design is increasingly being asked to ensure flexibility that can accommodate quantum computing. Gartner also recently released a report titled "Emerging Technologies for the Enterprise Data Center After 2030."
Gartner said design is needed with new computing platforms such as quantum computing in mind, not just AI. Because data centers are facilities intended for long-term operation, the company said they need to prepare for new computing environments that may emerge over the course of that operation.
The main force driving changes in data center infrastructure today is AI. As the number of AI-optimized servers increases, power requirements and cooling demands are rising as well.
Gartner projected that enterprise AI-optimized servers will grow at a compound annual rate of 13.8% from 2024 to 2030. Along with that, Gartner said enterprise data centers may need to support high-power density above 120kW per rack and liquid cooling in the future.
Quantum computing is being identified as a variable for data centers. The explanation is that, against the backdrop of several companies presenting quantum advantage and several companies having aggressive road maps to achieve FTQC around 2030, the timing for real business use is drawing near.
Quantum computing is expected to require a hybrid architecture combined with AI and HPC rather than standalone operation. Accordingly, heterogeneous computing system integration capabilities in future data centers are likely to become important, and Gartner judged that connectivity with existing computing systems needs to be considered. In addition, existing computing hardware such as GPU, AI ASIC, and FPGA can be used in the error correction, decoding, and algorithm support processes of quantum computing systems.
However, there is high uncertainty about the direction in which the technology and architecture will become established. Because it is difficult to predict which technologies and structures will become the standard, it is hard to design data centers tailored to a specific approach.
Quantum computing infrastructure has environmental requirements that differ from those of existing computing infrastructure, and one variable cited is that the mainstream qubit approach has not yet been clarified. Each mobility has its own design requirements, and those requirements include cryogenic requirements, vibration isolation, comprehensive Electomagnetic shielding, and specialized low-noise electronics.
Gartner noted that much of today's large-scale AI-optimized infrastructure is optimized for liquid cooling, which could create differences in cooling environments. The latest server infrastructure was built to use warm-water cooling, and the water temperature available on the "cold" side of the latest server infrastructure can be as high as 45 degrees. By contrast, the coolant temperature range required for quantum computing platforms is 15 to 25 degrees, so 45-degree water is not useful for quantum computing platforms.
Against this backdrop, the Gartner report said the need for AI, HPC, and hybrid architectures in quantum computing is becoming increasingly clear. Gartner said data center integration capabilities will be important going forward.
In the infrastructure industry, the need to consider quantum computing at the data center design stage is being emphasized. Because data centers are infrastructure with long-term operating characteristics, there is a growing view that they need to account not only for current AI demand but also for future computing environments.
In this context, Cisco is preparing for quantum computing threats. Cisco plans to phase in PQC algorithms based on U.S. NIST standards across the hardware and transport layers, and to have PQC built into new network equipment by 2026. It also plans to push ahead with a hybrid cryptography transition that meets the CNSA 2.0 standard by 2030.
Regarding this, a Cisco official said that Cryptographic Agility will become an essential requirement across future data center security. As a result, the entire encryption framework of existing infrastructure is expected to be rebuilt, which is being interpreted as a change aligned with a major turning point in the security sector.
Fortinet says that preparations are needed against threats to legacy cryptographic systems stemming from the development of quantum computing. Accordingly, it believes that post-quantum encryption should be reflected in data center design.
A Fortinet official explained that AIDC is infrastructure intended for long-term operation. The official also said that over the next 3 to 5 years, post-quantum encryption technology, taking into account quantum computing security threats, will become a core technology in AIDC SOC platforms.
However, it will be difficult in the short term for quantum computing infrastructure to become the main facility in general enterprise data centers. Gartner projected that the general enterprise deployment timing for quantum computing would be by 2030.
Initial deployments are expected to be centered on supercomputing centers, national research institutes, and some hyperscaler data centers offering quantum computing services. This suggests that the spread of quantum computing infrastructure will proceed around initial adoption sites with limited deployment.
Fortinet said that even before quantum computing is adopted, preparation is needed to accommodate related systems when considering long-term operation. The basis for this is that new computing platforms may be introduced during long-term operation, and the direction of this judgment is similar to the positions of global infrastructure companies.
Through its report, Gartner warned against excessive customization for designs centered on a specific computing technology, and emphasized the importance of securing design flexibility rather than making large-scale preemptive investments in a specific technology. The report described a trend in which designs should first be oriented toward accommodating diverse computing environments such as AI, HPC, and quantum computing.
Source: TECHWORLD · Kim Hye-jin
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406299
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.