What Quantum Computing May Mean For The Future Of IT Operations

Take a look at this post from Andreas Horn, Head of AIOps at IBM. Shared by Benjamin T.

I firmly agree that quantum computing's capability to tackle NP-hard problems, especially in complex operational environments, is promising.

While I believe that Large Language Models (LLMs) and Large Quantitive Models (LQMs) may and will get their "knickers firmly twisted" in operational environments, the potential of expanding operational causal models with Quantum powered counterfactual reasoning is highly promising.

Check out the posts from my friend Andrew Mallaband on Causal Models for operational environments. Causal Models capture operational behaviour independently of a specific environment. By instantiating these models with topology, a system computes a causal graph based on a Bayesian Network.

An abstraction layer, or "Codebook," then works out probabilistic Root Cause Analysis (RCA) derived from the Causal Graph. This is a complex but elegant solution to the incredibly challenging problem of defining root cause in complex environments, where services are layered across hundreds of thousands of combinations of underlying IT architecture components and micro-services.

Where I believe quantum computing may add value is in its ability to calculate Counterfactual Reasoning—essentially asking "what if" questions that define risks for any given operational environment.

Typically this refers to the exponential growth of possible states or configurations in a complex data systems architecture as the number of variables (N) increases.

With the scale that quantum computing could eventually achieve, it might become possible to identify the most significant service failure risks based on many (though probably not all) of the probabilities that could impact a given operational topology in advance. This is currently NP-hard, meaning it is nearly (if not completely) impossible to calculate all possible combinations of service failures in an environment with N to the power of possible combinations.

This is very promising indeed, though still some way off in the future. I would highly recommend anyone working in an operational environment to explore the work of companies like Causally and Causal Lens, and to at least take a cursory look into quantum computing and its potential for calculating probabilistic impacts.

The eventual combination of these technologies could significantly improve the operational management of highly complex infrastructures and may bring us closer to the truly predictive panacea of IT Ops, where problems are effectively spotted before they happen.

My interest in this area is primarily focused on operational data security architectures. Learning from and engaging with esteemed data security experts in the QSECDEF Community over the past few months—particularly regarding how quantum computing impacts not only cryptography but also secure communications—has been enlightening. Where I was highly skeptical just a few months ago, I have since "quaffed the Quantum Kool-Aid," and I am now a firm believer in Quantum's application, especially in defence, banking and other complex operational infrastructures.

Happy Sunday!