Autonomous vehicle delivery impossible without ‘extreme’ automation: Dell CTO

Because humans alone won't be capable of the processing power required to deliver future tech like autonomous vehicles, John Roese has said.

Our autonomous future: Driverless cars, drones, and robotic vehicles will drive unprecedented change and touch... Driverless cars are the world's most prominent example of AI and robotics in action. But driverless trucks, semis, delivery vehicles, and drones are poised to unleash a new level of automation in the enterprise.

Special Feature

Special Report: Autonomous Vehicles and the Enterprise (free PDF)

This ebook, based on the latest ZDNet / TechRepublic special feature, examines how driverless cars, trucks, semis, delivery vehicles, drones, and other UAVs are poised to unleash a new level of automation in the enterprise.

Read More

The year 2020 will mark the first step into the next data decade, according to president and CTO, products and operations at Dell Technologies John Roese, but as the world experiences an exponential increase in data, the software and applications realm is going to require a revolution.

"A lot of [that data] is being generated at the edge; but it's now clear that the topology of the modern enterprise is going to be a multi-cloud that's distributed spanning public clouds, SaaS services, cloud colocation, private data centres, mobile users, transformed buildings, and fundamentally smart everything," Roese said, speaking at the Dell Technologies Summit in Austin on Tuesday.

"What's interesting is that system is not just going to be idle, data is going to flow across the entire system in very dynamic ways, but beyond that, we in the IT community have created some very high expectations of this future. We have expected, or we created the expectation, that we're going to be connected everywhere with things like 5G."

The IT community, Roese said, has also created an expectation that AI-powered experiences are going to work on their behalf and make their lives better.

"We're all assuming that all this data and technology innovation is going to transform our businesses into productivity and disruption powerhouses," he said. "Now, to meet those expectations, which are very good expectations to have, we need more than just a multi-cloud system, and a new data ecosystem -- those are all important, but we're also going to need software and application revolutions to unlock the data value."

See also: Dell Tech Summit: Autonomous infrastructure, as-a-service models, and societal change (TechRepublic)

According to Roese, the limiting factor in unlocking this data era and this data decade Dell Technologies is betting so heavily on is not a technology or business model problem.

"There are plenty of those; the real problem is human capacity in the system," Roese continued.

"Today, the humans of the IT world, specifically developers, are still spending too much of their time dealing with low-level infrastructure tasks instead of building the algorithms and writing the code that are going to transform our businesses."

Roese said the best way for Dell to help solve this and make sure that it "isn't a catastrophic problem" is to change the way it builds platforms for digital business by "aggressively moving to automate IT infrastructure".

"I spent a lot of time over the last several years in the automotive industry, doing research and working with customers. The reason I'm here is not because of cars, it's because we believe those will be the first private zettascale infrastructures in the world," Roese said.

"What we learned are many things, but one of the biggest things that we learned was that without extreme levels of automation, it will not be possible to deliver autonomous vehicles, because we simply won't have the human capacity to get there."

One of Dell Technologies' global automotive customers by the end of the next decade expects to have 40 million autonomous vehicles in circulation.

"The dataset necessary to build out that intelligence system looks like it will be somewhere between 1-7 zettabytes under management inside of that one company," Roese explained. "Compute capacity is going to have to become cheaper, storage capacity is going to have to become cheaper, networking capacity is going to have to become cheaper -- but we can see a path to do that … I'm actually not that worried about those problems.

"But when we took a step back and looked at the human piece of the equation, and did the math, we looked at things like how many people would it take to run a storage environment of zettascale, and using best in class metrics today, like Isilon … we would need over a million storage units inside that environment just to make this real."

While Roese said clearly that's not possible, he said the shot clock is ticking towards the 2030 timeframe to improve the ratio of people to IT capacity by somewhere between three and five orders of magnitude before that point is reached.

"That's the challenge," he said. "We're seeing that challenge not just in automotive, we're seeing in healthcare, the 5G ecosystem, the financial services world, in fact, if you look at any industry long term and look far enough out, you will find a point that it is not the technology challenges that slow you down or stop you, it's the inability for us to scale the human capacity to actually operate these systems."

Roese pointed to consumer trust in autonomous cars, and said the data shows the world is not yet ready.

"We're at the stage now where automation can be aggressively implemented, but we have to take into account that there are human beings involved and they have to develop a trust for these types of environments -- and that's not trivial," he said, speaking with media.

Taking a look at the sequence of how autonomous cars and automotive innovation is occurring, the CTO said there is already fully-autonomous, completely self-driving vehicles, but they're operating in geo-fenced, very constrained environments like mines and industrial environments so they can be controlled.

"With the exception of Tesla that just does kind of progressive things very quickly, most of the mainstream automotive companies are in this phased progression, where the next phase they're entering is aggressive use of automation with one purpose and that is to make the people in the car safer," he added.

Highlighting Toyota's Guardian initiative -- the car-maker's big autonomous vehicle play -- Roese said the company's next step has nothing to do with driving a car, rather it has to do with looking at all the conditions in which a person could be injured in the act of driving, or sitting in a car, and de-risking it.

"That's their huge invention. Interestingly enough, in order to do it, you have to develop sensor technology, go and map the road -- all the things needed to drive the car, but they realised that the first step for an autonomous vehicle is creating a vehicle that just never crashes," he said.

"We want the technology to be a happy experience, a positive experience, safe experience, and everybody's now kind of institutionalised that … you don't get to the outcome if you don't bring humanity along with you, which is kind of exactly what's going to happen here."

Asha Barbaschow travelled to Dell Technologies Summit as a guest of Dell Technologies.

RELATED COVERAGE