Dell Technologies lays out AI strategy to unlock the power of data

Dell CEO Michael Dell and Dell EMC Global CTO John Roese laid out the Dell view of AI and the IoT, and the company's strategy to lead the market in addressing them.

Michael Dell, Dell Technologies Chairman and CEO, presenting at the ‘Unlock the Power of Data’ event

Last week, Dell Technologies held a virtual customer event, ‘Unlock the Power of Data,’ where the company laid out its view of the role of Artificial Intelligence [AI] and the Internet of Things [IoT] in changing business today, and how Dell is optimizing its strategy and products to best leverage it. The company laid out its view of the situation, and their very high-level strategy designed to meet it.

“It’s all about digital transformation,” Michael Dell, Dell Technologies’ Chairman and CEO said in kicking off the event. “AI and IoT are emerging concepts that are really changing our world.”

Dell compared the impact of AI to Gutenberg’s invention of the printing press in the 16th century.

“The invention of the printing press, part of the disruption of the Renaissance, was the birth of information technology,” Dell said. “We are now standing at the cusp of another Renaissance with the advent of AI, broadly referred to as the digital transformation.”

Dell cited data indicating that 67 per cent of G2000 companies have digital transformation as a top strategy, and that 7 of the 10 most valuable companies in the world are using deep learning and AI at the heart of their operations, to deepen customer relationships, grow new capabilities or design better products.

“Technology strategy is now business strategy,” Dell stated.

Data, on the other hand, helps make a product or a service better, he emphasized.

“The big telcos are putting in 5G networks not so you can talk on the phone faster, but for more data,” he said. “If AI is your rocket ship, data is the fuel for your rocket. But not all data is good data. It requires data engineering, and when you get it right, it becomes data capital, your most valuable asset.”

Dell said that engineering requires a new type of tech infrastructure that seamlessly embeds AI and IoT, from the edge to the core to the cloud. It involves a highly distributed AI infrastructure at the edge, through to the data centre, with highly automated and intelligent cloud infrastructures. It also involves partnering with all of the public clouds, and Dell Technologies’ own specialized cloud capabilities with Boomi, Virtustream and Pivotal.

“It is a holistic approach to the whole system,” Dell said. “It is why we believe Dell Technologies is best positioned to simplify AI so customers can participate in the data revolution. We are the platform for AI, the platform for human progress.” They are not about nichey high-end applications – what he called the last vertical foot of AI.

Dell summed up his presentation with three thoughts.

“These are emerging and powerful technologies and are bringing us into the next Renaissance,” he said. “These new technologies will require a massive investment in new types of AI to enable infrastructures to seamlessly connect algorithms from the edge to the core to the cloud. This is why Dell Technologies was created, to take the opportunity head-on for our customers, to be the platform of choice for AI and human progress.”

John Roese, Global Chief Technology Officer at Dell EMC, provided further high-level detailing of Dell’s strategy in the space.

“In the Big Data era, most of the data we collected was ‘write once, read never,’” he said. “The data went to die in a data lake.”

That has all changed in this new era of data, Roese stressed.

“We will see a hundred-fold increase in the amount of data that goes through analytics yearly,” he said. “It’s an entirely different era of data because of the scale and capabilities. Investments in digital transformation are growing at 18 per cent year-over-year. This era of data is about focusing on data to achieve digital transformation, not just having a lot of data.”

AI and machine learning are the tools to unlock the data, and Roese said it would lead to major transformation in three ways.

“We will transform the user experience,” he said. “Today, most interactions are passive and static, but when you put machine intelligence and a stream of data behind it, they become interactive. Even simple web apps powered by machine learning give you a totally different experience, which becomes hyper-personalized.”

The second change is that machines will do more and humans less.

“Every business process that matters will apply machine intelligence to change the line between tasks done by humans and by machines,” Roese said. “That line will go up, with more being done by machines.” He said the results would be arguably the latest potential productivity gains ever.

“We expect upwards of 6 billion person-hours will be absorbed into infrastructure, and those people can be repurposed to things that create new value,” he stated.

“Thirdly, in this data era, these tools will give us capabilities to do things we could not previously dream of doing, by being able to shift the entire burden of tasks onto the machines,” Roese added.

The building blocks of this will also shift. The data will be more specific to particular enterprises. So will the algorithms and applications.

“In the last 20 years, these have been general purpose,” Roese said. “We all had the same CRM system. In this data era enterprises will build their own, with things that are unique to them.

The infrastructure underneath is the final component, and that is critical to Dell Technologies.

“That’s our business,” Roese said. “We are incredibly focused on providing that infrastructure to the IT world.”

The edge is a critical part of this whole process, and Roese stressed that it’s not just an accessory.

“It is as significant as the private data centre and the cloud even though we are just getting started,” he said. “People are underestimating how many things will consume and produce data at the edge. The total today is only about only 150 million devices, but it will grow exponentially year over year. We are also underestimating the connectivity of that edge. 4G is a great technology, but it has been a bottleneck here. 5G will expand that exponentially. It also pushes a compute layer to the edge. That’s an incredibly powerful tool. Finally, we are underestimating the amount of data these things will produce. Something like connected cars will produce MB of data per second. There are huge insights to be gained by having sensors in that environment.”

So how does Dell Technologies take advantage of these trends? Roese said that there are six dimensions of innovation that the company intends to pursue.

“First and most important, we must have a new processing paradigm to unlock the data,” he stated. “It is general purpose processors complemented by diverse accelerators designed to optimize specific machine intelligence tasks – diverse compute. We have seen that already on Intel road maps and startups in this space. It’s necessary to keep up with the processing demands of these technologies. Our recent new PowerEdge servers were dense accelerator servers as a result, with a lot of ports and slots for accelerators. That is a critical architectural shift.”

The second focus area is edge compute and analytics.

“We believe our market leadership in data centre servers and compute  will let us build out at the edge – even though it’s not the same type of compute,” Roese said. “It still requires deep computing expertise, so our heritage in data centres allows us to dominate.”

Storage and data protection are changing to meet these trends, Roese stated.

“Until now, storage has been an OR game – high performance storage system OR high capacity OR high resiliency,” he said. “In the data era, it’s an AND game.” All of them are needed. Many Dell EMC products have reflected this in their new upgrade cycle.

“When we introduced an all-flash Isilon, many thought it was overkill, but you need massive throughputs to pull in the data, and all-flash competes that,” he said.

Multi-cloud operating models are the fourth focus area.

“We recognized the future would be hybrid and multi-cloud, and it was our job to create a uniform experience cross them – edge to core to cloud,” Roese said. “Having software and cloud architectures to let you operate that way is critical, to make sure the technology can work in a multi-cloud context.”

Software-defined infrastructure is of critical importance.

“This does not mean ‘Hardware Bad and Software Good’,” Roese noted. “It means we have no idea what demand loads will be, so infrastructure has to be agile and making it behave like a programmable software architecture is the best way to do that, with reconfigurability and  programmability of systems.”

The sixth focus area is data mobility.

“This is because of data gravity,” Roese said. “Moving an exabyte across the Internet takes a long time. Knowing where your data is and being able to move it efficiently, or having it in right place to begin with, is critical. We are in a good place here, with our leadership in synchronous and asynchronous replication. Boomi also makes sure data is synchronized before it’s processed.

“If we get these right, we have abundant compute, and can extend it to the edge and can make infrastructure as agile as the applications above,” Roese concluded.

Referencing an earlier presentation by Nick Curcuru, Vice President, Data and Analytics at Dell customer Mastercard, Roese noted that Mastercard is already on this journey.

“They are already using all the tools we talked about,” he said. “Caterpillar is another good example. The real implications of the data era isn’t the technology. It’s using it to do actual things. With health care, we can finally achieve that reality we’ve been talking about for ten years, where an EMT in an ambulance can use augmented reality to share data dynamically with a physician in an emergency room. We’ve been talking about it for years, but we didn’t have the tools to get there. We’re going to get there.”