For more than two decades now, quantum computing has regularly appeared in the news headlines as a potential new wave of information technology.
Since 2019, we have seen claims of ‘quantum supremacy’ where a quantum computer is said to have performed feats impossible for the conventional equivalent. The first such announcement was Google’s claim to have taken just 200 seconds to perform operations that would take a supercomputer 10,000 years to complete. But how much is hype and how much reality? And should we be getting ready to replace our conventional machines?
Rather than work with the familiar ‘bit’ at the heart of every conventional computer, which can hold a value of either 0 or 1, the quantum computer uses ‘qubits’. These are quantum particles, such as electrons or photons of light, which in effect can handle any value between 0 and 1, and which can act together to multiply up their efficiency.
There is no doubt of the potential power of quantum computing – but there are problems too. Qubits often need to be kept in special conditions, such as near to absolute zero (-273.15 °C) to maintain their values, making them anything but desktop. It is also not a trivial task to get information into and out of a quantum computer without errors being introduced. Most importantly, quantum computers are not universal. While there are specific algorithms that will run on them (some developed decades ago), they will never be general purpose.
As yet, claims of quantum supremacy need to be examined sceptically. In part this is due to the difficulty of making comparisons. As soon as Google made their original claim, IBM suggested that rather than 10,000 years, the operations would take only a couple of days to undertake on (their) supercomputer. But more significantly, when such claims are made, they usually involve a specialist task that does not have any relationship to the real world. There is indeed supremacy – but at performing a task no one wants to execute.
When coders do start to employ quantum computing for business purposes, it’s likely that two of the earliest quantum algorithms to be developed will be at the heart of their use – one of immense value when performing searches, the other putting internet security at risk.
The first of these was the work of Lov Grover, then at Bell Labs in America, first appearing in a Physics Review Letters paper titled Quantum Mechanics Helps in Searching for a Needle in a Haystack. If you imagine writing code to search for a particular piece of information, which could be in any one of a million locations, in principle you might only find it on the millionth try – on average it would take 500,000 attempts. Of course, modern search technology provides many mechanisms to get around this. But for unstructured data with no indexing this is still the case. However, Grover’s quantum search algorithm guarantees finding the result in the square root of the number of locations. That’s just 1,000 tries in the example above.
The other classic quantum algorithm was the work of Peter Shor (also, at the time, at Bell Labs). This solves a tricky maths problem with a number of potential applications in computer science. Some mathematical processes are difficult to reverse. For example, if you multiply two extremely large prime numbers together, it can take a vast amount of computing time to work out what those numbers were from the result. Shor’s algorithm makes this process far quicker.
Unpicking such a multiplication has the potential to help with a classic difficulty faced by satellite navigation systems: the so-called travelling salesman problem. It rapidly becomes impossible to work out the optimum route from A to B as you add in more roads, more connections and complications like traffic delays. Although modern sat-nav technology does a good job, it doesn’t provide the best route – just an adequate one.
The kind of mathematical leverage provided by Shor’s algorithm could help fix the travelling salesman problem.
But unfortunately, it also has the potential to break RSA, the security technique used widely on the internet. It is hoped that by the time quantum computers are robust enough to take on this challenge, better security methods will be widespread – but it is a concern.
The reality, then, is that you will not get a quantum computer on your desktop. They are likely to remain as specialist devices running from data centres. While they may become central to functions such as search engines and navigation, both cloud-based, for most coding they are likely to be occasionally accessed remotely to undertake a specific process before returning results to a local device. It’s true that quantum computing could transform search – and threaten current internet security – but the claims of supremacy should be taken with a large dose of salt.