By Richard Murray
We are close to a commercially useful quantum computer, which raises a significant question: How will it integrate with conventional data storage?
As systems, the two represent fundamentally different technologies that must be made to work together. But little work has been done to ensure such harmonisation happens.
In fact, it is striking how often quantum computers and quantum communications are considered as if they will exist in hermetic isolation, which is an unlikely outcome given the general direction towards co-ordinated, integrated systems. It could prove a costly belief.
Take a look elsewhere. There may be many ways to improve trains as pieces of engineering, for example. But unless the result can run on the standard gauge, which is nearly 200 years old (and based on standards set for Roman chariots) those improvements will quickly hit the buffers.
The same principle applies to quantum computers: If they cannot work with the existing 8.3 million data centres worldwide then their value will be severely limited. Not that the task of integration will be easy: Quantum technologies are fragile enough in their own environment without trying to connect them to others.
Part of the omission in thinking through what the new quantum era will require is because quantum technologies are being developed in isolated bubbles. These are failing to consider the requirements of data centres and their users to make sure it all works efficiently. The result of this isolation is that quantum advances in computing and communications are just not on any course to readily be adopted or integrated into existing infrastructures.
This oversight is already painfully apparent. Whilst Google, IBM and Rigetti are technically advanced with 50-plus qubit platforms, they are all without an engineering framework for real-world deployment. In quantum communications, the market leaders have also failed to develop significant links to the data centre community.
This core challenge of integration must be addressed. Data centres are what allow our increasingly digital economy and society to function. They are its workhorse, driving growth and underpinning our lives. The demands on them are growing, not least as Artificial Intelligence (AI) adds complex workloads, making centres reliant on innovative hardware and architectures to keep pace.
There is another reason to act now. Increasingly, data centres are being used to perform challenging calculations, not just store information. This has led to a rush for ever more efficient architectures that can use a range of processors, which itself makes the case of quantum processors that are able to process deeper, complex data sets
When quantum computers do arrive for commercial use, probably as hybrids working with existing supercomputers, they will have to work with the way the digital world is ordered now, not how it might be at some future point.
The challenge of integration will be considerable, but must be addressed. We are already producing an extraordinary amount of data even without the enormous information generating power of quantum, much of which will need to be accessed into the future, including the 188 million emails sent every minute. Some estimates suggest that by 2025 the amount of daily data generated will be 463 exabytes, which is the equivalent to 212,765,957 DVDs every 24 hours.
Those who expect that quantum will itself be an integrated world, a closed loop of its own bespoke technologies are gambling. Perhaps one day. But experience suggests that this is unlikely in the medium term.
The history of industrial progress is messier and more incremental than clean breaks between what was and what has arrived, something which has as much to do with non-technological drivers, such as cost and conservatism, as it does with any enabling value. The dominance, let alone ubiquity, of a technology usually comes over time. In the period between breakthrough and predominance is a time of accommodation and collaboration. There is no reason to suppose quantum 2.0 will be any different.
Richard Murray is CEO and co-founder of ORCA Computing.