I'm a fan of CES, and have attended regularly for over a decade. I especially like to focus on the young companies who, sometimes, promise to disrupt the market.
The disruptive tech triad
Today the three most important developing technologies are artificial intelligence (AI), blockchain, and Internet of Things (IoT). All three are highly centralized, relying on massive computes and datastores for value exploitation.
Of course some data needs to be aggregated, and the cloud is often a logical place for that. But my question is how much processing can be done locally to reduce the network overhead and limit the leverage cloud providers have over these apps and, more importantly, us.
As we have seen with Facebook and other large cloud operations, corporations are more concerned with making money than they are with information and application security. Which is why I worry about these disruptive technologies relying on cloud vendors.
What could possibly go wrong?
Amazon, for example, hosts storefronts for numerous small vendors, but it also sells Amazon-branded products that compete with products from those same vendors. Does Amazon use the sales data from its customers to decide what it should offer? I don't know, but it wouldn't surprise me.
Or what about cloud-based blockchain services? Even if the content of the blockchain is fully encrypted, the use of signal intelligence could enable the cloud provider to determine who the parties are and the likely nature of the transactions. Arbitrage or stock market plays are two obvious exploits, and I'm sure devious people could think of many more.
Those are things that cloud providers could do on their own. But what happens when they are compromised by state actors or criminal groups?
The Storage Bits take
The fix is to have an alternative to the cloud vendors. Warehouse scale computers are the 21st century's answer to 20th century mainframes. And the answer to those was distributed computing.
Today, that means moving computing to the edge. Is that possible? Sure. Easy? No.
But as these applications are better understood, we can create silicon that enables local processing at the edge. With 5G bandwidth, terabyte phones, and, say, convolutional neural processors, we could have mini-clouds at the edge.
To make such things cheap enough means expensive R&D and huge unit sales - at least 100 million or more. While the giant cloud vendors may be part of the problem, only the largest hardware vendors will be able to provide a local computing solution.
Courteous comments welcome, of course.