Quantum Computing and Data Centres: The Potential Impact of Quantum Computing on Data Centre Infrastructure
Sudhanshu Mittal, Head & Director Technical Solutions, nasscom MeitY CoE– IoT & AI discusses how quantum computing can revolutionise data centres by handling complex tasks and cutting power use. He highlights its potential in various fields, including fighting deep fakes. Mittal also addresses challenges like integration and data security, suggesting a future where quantum and regular computers work together. The shift could make data centres more accessible but requires professionals to adapt to managing both types of computers effectively.
What are the fundamental principles that make quantum computing a game-changer for data centre infrastructure?
Quantum computing has a strong role to play in the data centres. One of the key requirements for data centres is the ability to process complex algorithms being used by the developers. Quantum computing, with its extremely high processing capabilities is ideally suited to address the challenges in processing such complex algorithms.
Another critical impact is in terms of the power consumption. Data centres are consumers of huge compute power due to very high computation requirements which require large number of high energy CPU/GPUs. Quantum computer’s capability of providing extremely high processing capability without consuming very high power is the ideal match to reduce the overall power consumption of data centres – both for addressing the high processing requirement and the attendant cooling requirements to address the heat generated.
Quantum computing is often associated with its potential to solve complex problems. How do you envision quantum computing being applied within data centre infrastructure? Are there specific use cases or industries where quantum computing could have a significant impact?
All kinds of use cases involving complex algorithms are the candidates for being addressed using quantum computers. Use cases around financial modelling and risk analysis at macro level, environmental analysis, and climate modelling especially while undertaking the development projects in ecological sensitive areas, supply chain optimisation, life sciences, AI based drug discovery / drug repurposing, custom treatment for complex disease etc would be the candidates for using quantum computing in data centres.
Apart from the above, one key area where quantum computing will impact everybody is deep fake. In a very short time, the generative AI has shown its capability of creating fake videos of anybody, with little training material. Identifying such deep fake on social media platforms is an extremely critical challenge as the required algorithms are complex and need to run on millions of the videos floating around to identify few fake videos. Quantum computing will play key role in providing the required infrastructure support implementation of algorithms which will be able to identify such fake videos and stop them before they become viral and create law & order problems in societies. Players like Facebook (including WhatsApp), Instagram have strong usage requirements of quantum computing to address the menace of fake news and fake videos.
Quantum computing is still in its early stages. What are the challenges and obstacles that data centres may face when integrating quantum computing into their infrastructure? How can they navigate these challenges effectively?
It is expected that quantum computers will not fully replace but will work in tandem with classical computers. When we talk of data centres, there has been huge effort spent in building the software stacks and integrating those to quantum computers will face significant challenges. Quantum computers have their own strength in processing capabilities while classical computers are general purpose computers, capable of addressing a large variety of problems. Ultimately it is expected that we will have a hybrid model where classical computers will be used to manage the quantum computers by offloading the specific parts of problems to quantum computers while doing post processing themselves. While these are early days to make definitive predictions, it may work in the manner of current CPU and GPU model where CPU is used for overall activities while offloading the specific computations to GPU.
Data security is a critical concern for data centres. How might quantum computing impact data security and encryption practices? Are there new approaches or risks that data centres should be aware of in a quantum computing era?
Traditional data security is based upon encryption using secret keys and complex mathematical algorithms. While in theory any encryption can be broken using brute force approach, the estimated time taken of 1000s of years makes such proposition meaningless as by the time data is successfully encrypted, its usefulness will be long over.
The processing capabilities leap generated by quantum computers threatens to upend this comfort factor of data encryption. While it is still early days, it is being estimated that quantum computers will be able to break the RSA 2048 key within a few years (in single digit), thereby giving access to data while it is still relevant. Imagine the specifics of military engagement between two nations – such engagements need to be kept confidential for 20 years or more, in some cases much higher. If using quantum computing the encrypted data is broken within two years, the adversary will have access to information while it is still relevant and in operation.
As quantum mechanics provides the capability to break the traditional encryption, it also provides the ability to protect data in the quantum era. The Quantum Key Distribution (QKD) is the methodology that uses the quantum mechanics principles to enable data communication while guarding against external snooping. It will be equally critical for data centres to integrate the QKD capabilities also along with the integration of quantum computing.
Looking ahead, what are your expectations for the evolution of quantum computing in data centre infrastructure? How might it reshape the landscape of data centres in the coming years, and what should data centre professionals prepare for?
Quantum computing heralds the potential to make data centres much more useful and relevant for users, thereby democratising the access to quantum computers for the masses. The complex algorithms which require access to supercomputers currently, will be able to run on such hybrid data centres very easily and can be easily used by normal users who have basic accounts at the data centres (or even free educational credits).
It may be necessary for the users to learn how to make effective use of the processing capabilities of quantum computers, instead of just relying on the classic computer component of the data centre. As for those professionals managing the data centres, they will need to also learn how to manage this interoperability between classical computer and quantum computer apart from having to maintain the huge stack on quantum computer, thereby making their life more challenging and interesting.