Monthly Archives: December 2011

Time to start investing in quantum computing software?

While the argument over whether D-Wave represents true quantum computing still rages the company moves on and opens a software developer portal.

Assuming D-Wave succeeds in the marketplace (and I wish them all the best) they will own the term “quantum computing” for the time being. It is very unlikely that another vendor will introduce another quantum computing device in the intermediate future. Being the ultimate first mover in this nascent market D-Wave has the unique chance to shape it to their advantage.

The (mostly academic) critics won’t cease to point out that D-Wave doesn’t implement a universal gate programmable quantum computer – not that the company ever claimed such a thing – but assuming that their device delivers superior computational power the market won’t care.

But there is a catch: D-Wave’s optimization engine is versatile enough to help with many real life optimization business cases but the path to actual usefulness is all in the software.

Let’s first look at the B2B market (I’ll discuss applications in the consumer space in a later blog entry).

It is hard to get market share numbers that break out spending on IT solutions for optimization software. Currently this is mostly folded into the BI and CRM category, both areas that pose plenty of business cases (e.g. channel, price, revenue, supply chain optimization to just name a few). According to Forester, BI and CRM are the main drivers of IT spending in 2012.

Even if D-Wave could just dip a toe into this huge market it’ll render it a commercial success.

But selling any optimization solution to businesses is no small feat. The potential customers don’t care about the underlying technology and won’t be interested in talking about hardware and algorithms.

Making a persuasive case is tricky because it requires a special skill mix: Domain expertise as well as understanding of the parameters within which an optimal solution can be found. While it suffices to mention this as “boundary condition” to those with a mathematical background, this term will result in blank stares with your average business crowd and is guaranteed to sink your sales pitch.

The magic realm of financial engineering may seem like a good fit for D-Wave. The quants who reside in this world will clearly understand where this company is coming from – many of these folks will have been lured away from a career in math and physics and there is no doubt that they could utilize a quantum optimization device to its fullest potential. That is, if this magic kingdom wasn’t on fire. Wall Street’s profits are up again, but on a shaky foundation, and the faith in the reliability of financial modeling has certainly taken quite a hit. (For a great account on this I recommend the hilariously named “Models behaving badly” book). This will certainly have reduced the appetite to gamble sizable amounts on – as of yet – untested technology (although sizable is of course rather relative when contrasted with the kind of volume that high frequency trading employs).

In the end, D-Wave’s best option may be to lower the barrier for engaging with their tech, i.e.  by developing a stack of accelerator software that allows programs written in SAS or R to execute without modification in a hosted environment. That means engineering a pre-processor that diverts the invocations of proc nlp or optim() to an underlying D-Wave parser that compiles the optimization problems for execution on their hardware.

“Build it and they will come” is rarely a winning business strategy, but if D-Wave can make it effortless enough to test drive their optimization engine, then they stand a good chance to competing on the merits of their technology.

When they finally came ...