There have been numerous posts and tweets coming from the NonStop vendor community following RUG events worldwide: ETBC,...
Quant … ummm? Is it real?
I was taken by surprise at this week’s SIBOS Sydney 2018 event. This is a major conference for the financial industry, with numerous high-profile banking CEOs participating in the panels and round table sessions. While I had heard about SIBOS this is the first time I have participated and I have to admit, even as jaded as I am at times from the many events I attend each year, SIBOS Sydney 2018 is delivering the good on all fronts, serving up relevant topics for bankers and technologists alike, but Quantum computing?
Surely I had misread the program, but no, for almost the entire day, at the SWIFT Innotribe Theater it was the topic of the day. Located on the upper levels of the exhibition hall, the introduction music as each session on Quantum began was enough to end all conversations on other stands throughout the entire hall. As the day progressed the topic moved from a simple explanation of Quantum on to why it should be considered and then followed what it is doing today and in all likelihood, tomorrow and yes, even a demonstration.
I am sure we have all heard something about Quantum physics and Quantum computing but I am not sure it really was on the radar for most banking executives. However, following the presentations I can clearly state that it should be as this is an area of computing that is moving very rapidly. The driving force behind all of this interest is the acknowledgment that Moore’s Law is coming to an end – there is just no practical way to compress even more transistors on a square millimeter of silicon. It’s over and the industry has to turn to other options if we are to handle the volume and velocity of data heading towards the data center.
It’s almost by coincidence that we are reaching the limits of silicon just as we are getting ready to face down the full impact of a connected world where everything computes and everything communicates. The Internet of Things (IoT) is about to make its presence felt even in traditional industries such as banking, but moving beyond silicon and the traditional model of processor, memory and storage hierarchies along with networking is a must and vendors of all sizes are looking at potential alternatives to the traditional model. For many years now, HPE has been talking about memory-driven computing solution, The Machine.
Memory-Driven Computing? I recently posted about memory driven computing following remarks on this subject by HPE CEO, Antonio Neri. Yes, we are seeing it being introduced into the marketplace already as the SuperDome Flex – it’s definitely a memory-driven computer and it is definitely foreshadowing the potential that is an extension of the first prototypes coming from The Machine. But wait, there’s more – Memory Driven Computing is at the very heart of the converged and hyper-converged models HPE foresees the market requiring to process the workloads that will arise from all the data being ingested.
When you look at the charts HPE is showing how this new model for computing is being developed – with terabytes and yes, projected petabytes of main memory (and no storage hierarchy whatsoever) – processors simply “plug in” and among the processors that are being targeted by HPE to plug in, is a quantum processor. At this SIBOS Sydney 2018 event I was enthralled to hear that the initial approaches were all hybrids in nature – indeed, the demonstration given in the afternoon was a hybrid of quantum and CPU/GPU chips.
And this simply makes sense. But there is also another important aspect of quantum computing – the use of quantum to model quantum. At the recent HPE Discover event in Las Vegas, HPE Senior Fellow, Ray Beausoliel, leading the team researching “beyond transistors” and yes, even “beyond qubits” has been working with quantum computing. The presentation, Do we have to wait for Quantum Computers followed along similar lines to what was presented at SIBOS and covered a lot of territory before concluding that as “quantum computing influenced new ways to process information, it wouldn’t be used on current applications but rather, it would teach us something we didn’t already know,” said Beausoleil.
In other words, apart from “quantum computers being really good at modelling other quantum systems,” and if we want to know more about what’s next after silicon should we want to extend Moore’s Law, then “quantum is good for figuring out what’s next; consider a petabyte being stored on 10 qubits!” Pretty much the very same observation was made by one of the presenters this week at SIBOS. HPE is bringing The Machine to market with SuperDome Flex being a working model you can buy today. But as The Machine continues to mature, it is clear HPE plans to add something very special when it comes to quantum computing.
And this proved to be the common theme at SIBOS. Other presentations that continued throughout the day talked along similar lines and when you see other mainstream vendors like IBM actively pursuing quantum computing you have to believe that there is more than vaporware in play here. It may not be real just yet but we are oh so close to seeing something we can begin taking very seriously very soon.
Of course, listening to the music thumping across the exhibition hall at SIBOS proved to be a drawcard for many and the sessions were very well attended. But as excited as the industry is becoming over quantum computing (even as the end of Moore’s Law shocked others), and we look beyond simple use-cases with respect to security and cracking complex primary numbers, etc. – will bankers warm to spending the types of money being projected for early deployments of quantum computing?
In all seriousness, it’s all a bit too much. It’s overwhelming in its complexity and it’s still the closest that technology has come to magic – quantum entanglement, anyone? On the other hand, as the first “magic” is put to use and solutions are produced then like no other industry I know of, there will be a rush by all financial institutions to follow suit once any one bank jumps on quantum and solves a real business problem. It will happen and when it does in the not too distant future, there will not be anyone left yawning over this topic. Quant … umm? It could very well prove to become Quant … yumm!
[Reposted following posting to the LinkedIn blog, Pulse on October 24, 2018]