I am in Guadalajara, Mexico, the big storm never got here, in spite of doomsayers. I am giving a keynote on Innovation in Smart Cities of the future and I share with you a few thoughts in this and the following posts.
I start by looking at technology evolution since technology is playing a significant part in the economics of cities "smartification". According to ABI Research 39.5 B$ will be spent worldwide in 2016 on technologies to make cities smarter.
The relentless evolution of processing capacity, clocked by the Moore’s law, is now changing shape. Further shrinking of silicon is no longer possible mostly because of economic hurdles. For the first time in 2014 the cost per transistor in chips have started to grow bringing the economic side of Moore’s law to a screeching halt. Someone has pointed out that this may well result in a changing of the business landscape in this area. We might go back to the time that upgrades sustained a thriving industry. That industry was killed by the short life cycle that brought a new, more performing release before any upgrade was able to generate returns.
There are, however, alternatives to silicon: the so called 2D substrata (graphene, molybdenum disulfide, black phosphorus….).
Even more important, evolution has not halted in terms of power savvy and we might expect very low power devices that can feed on scavenging. This will boost sensors, will reshape our perception of many objects in the urban landscape.
We are surely not near the end of processing application and leverage. Again, the urban landscape is going to be significantly reshaped because of the capillary presence of processing anywhere.
Storage has increased in density, we have reached in August 2015 the 16TB on a single SSD disk (Samsung). This advance has been made possible by the V-NAND (where V stands for Vertical) architecture, layering one layer above the other of transistors, hence packaging much more storage capacity on a single chip. The 16TB SSD uses V-NAND chips with 48 layers.
Intel and Micro 3D Xpoint are working on a completely different approach to storage getting rid of transistors. The storage points are created by mesh of wires that can be stacked. This results in a 1,000 fold speed increase and a 10 fold density increase. Expect to see this in a new generation of chips later in 2016.
More disruptive performance leaps can be expected by radically new approaches, one of the most interesting based on spintronics. Skyrmions are part of this new wave of radically different ways to storage. They use localized magnetic fields requiring extremely low power.
Interestingly we have seen progress in storage also at the very low end of performance, like in QR codes. The interest is in their very low cost and easiness to read (given the spread of smart phones).
At a city level this means that, more and more, any object can contain data that are easy to capture. Think in terms of replacing all labels on roads, cabinets, benches… whatever, with relatively small quantity of data. Actually, most of these data can be pointer to places in the web where the full slate of information is available.
A smart city should have data associated to every single object that is deployed.
Why have I been talking about progress at the high edge of performance for both processing and storage when later I said that the interest for smart cities lies most in the low power consumption (processing) and in the QR/tags (storage)? The point is that progress at the leading edge is what is and will be transforming smart phones, cars, buses... This will be crucial for Smart Cities because they can leverage on this capacity at the edges without having to invest to create such a capacity. Of course, to do that we need a pervasive communications infrastructure reaching these "terminals" and having them becoming part of the city overall infrastructure.
That is what I will consider in the next post.