Quantum Computing - not just a security issue

The Zeitgeist

A quick Google search for Quantum Computing seems to suggest that the current focus of efforts can be grouped into three broad categories:

• The Security Threat of Quantum Computation: threats are perceived, such as using quantum algorithms to speed up the process of factoring semi-primes, a key process for cracking RSA public-key encryption (Cybersecurity Threat). Alongside this research is attempting to find “quantum-safe” encryption methods.
• Combining Quantum Computer techniques with AI and ML: AI/ML models are being trained with increasingly large data sets, so much so that the carbon footprints of some models are becoming concerning. Quantum algorithms promise the ability to speed up some AI/ML algorithms, or even make possible algorithms which are computationally hard (Quantum AI).
• Quantum Supremacy: Quantum supremacy is defined as the moment when a quantum device is able to execute an algorithm that would take a classical computing device an infeasible amount of time to solve. According to Google they have already achieved this back in 2019 (Quantum Supremacy).

Whilst each is interesting in its own right, and the security implications is of critical importance for organizations, little is said on the more mundane uses of Quantum Computation. Do Quantum Computers have a place in more general applications, for example ERP systems.

NP-Complete and BQP

Classical computing divides the space of all problems (known as PSPACE) into separate areas of differing complexity, and increasing difficulty:

• P: Problems which can be efficiently solved by classical computation techniques.
• NP: Problems that cannot be solved efficiently using classical computation, but if a solution is given it can be verified efficiently.
• NP-Complete: More complex NP problems, which every NP problem can be transformed into in polynomial time. This means, if we can solve an NP-Complete problem efficiently, than we can solve all NP problems efficiently.

This leads us to quantum algorithms. BGP (Bounded-error Quantum Polynomial Time) problems are those where a quantum algorithm can be found which will find a solution in an efficient time with a small chance of error.

Currently it appears that BQP problems will encompass the P problems of classical computation, along with some of the NP problems. It is an open question whether an efficient quantum algorithm can be found for an NP-Complete problem.

What Does This Mean?

What this all means, however, is that quantum algorithms can be developed to solve currently unsolvable problems, not just to break encryption or speed up AI/ML algorithms, but also problems which have a practical application in business. For example, the Knapsack problem (NP-Complete) which is a generalization of resource allocation problem such as found in project management, ERP, or stock portfolio optimisation.

Quantum Computing will fundamentally change the computational landscape, eventually. At the moment algorithms are limited, due to the small number of qubits which can be implemented in a quantum computer. However, given the current pace of development Quantum Computers will become useful in a short span of years.

Therefore, business should be starting to think about a quantum strategy, not just how to be secure in a post-quantum cryptography world, but also how can quantum algorithms be used in their systems.

What this could mean for Architects, be they Enterprise, Application, Data, &c. has been covered in an excellent article by Redhat (Quantum Architect). Of key concern for me, as a Data Architect, is how do we deal with data that is probabilistic and no longer precise.

#quantumcomputing #enterprisearchitecture

1 Photo by Farai Gandiya on Unsplash