2020 was a year of acceleration, with what used to take five years being done in a year or less, observed Dell Technologies' Amit Midha, who also spoke of having a front-row seat to overnight digital transformation.
"We want to be the essential infrastructure company for the data era," he said, pointing out that the company's PowerScale, PowerStore and Project APEX launches in 2020 have been the instrumental to enabling the remote work lifestyle that the pandemic has encouraged.
John Roese, Global CTO, Dell Technologies, added that Dell Technologies is uniquely positioned to win. "We've done relatively well in this very complex time because technology still matters," he said. He noted that the client solutions group had seen record shipments in the past quarter, an indication of demand for compute in the home, especially when each person in the household now needed their own device.
 |
| Source: Dell Technologies. Last year was about helping customers win, said Roese. |
"Last year was absolutely about helping our customers," he said. "It wasn't about the technology or the business."
Roese also shared several 'less visible' technology predictions for 2021:
5G
2020 was the first year of 5G, with networks extending 4G (non-standalone 5G) launched in many countries. True standalone 5G is the game-changer, bringing initial use cases beyond slightly faster consumer broadband to the enterprise, Roese said. "It will shift to an enterprise focus, revolutionising healthcare or logistics. A secondary thread will be faster broadband for people to watch YouTube," he suggested.
5G players will also include non-telco companies such as Dell, Corning and Microsoft, as 5G architecture evolves from the traditional telecommunications architecture to become more cloud-centric, open, software-defined, and standardised.
"That's why Dell is involved," he explained.
Edge
Edge computing is in trouble as "everybody built an edge", Roese said, noting that it leads to compatibility problems.
"There are too many independent edge systems – we will have a proliferation of independent, siloed edges," he predicted.
Roese's view is that every edge experience in 2021 is an independent infrastructure, but that a shift will occur towards 2022 where some players will shift to building edge platforms that run multiple edge workloads as the model is more manageable and cost-effective.
He said: "The edge discussion will become two discussions – a discussion about platforms and pools of shared capacity, and (a discussion about) edge workloads – extensions of processing and data tasks."
Players in the edge space will therefore either deliver workloads or build platforms that run those workloads, Roese said.
"Dell is very focused on those platforms," he disclosed. "There'll be a limited number of edge platforms, but dozens of edge workloads that run on those platforms."
And while edge is typically taken to mean devices on-premise, Roese pointed out that edge functionality can be delivered as a service. This could mean a hospital or factory can avoid rearchitecting infrastructure just to adopt edge computing.
"They can deploy edge capability without having to rearchitect it themselves, and get the benefits," he said.
Further, public cloud and SaaS edges will become software-defined workloads and depend on edge shared platforms, while edge platforms will become major new areas of on-premise IT capacity that is delivered as both a product and as-a-service.
Another implication of edge computing is around the need for domain expertise and integration capability.
"Platforms and technology are important but the ability to get them into production requires expertise," Roese said.
"Our edge strategy is extremely focused on deep partnerships because the outcome is not building an edge."
Quantum
This prediction would not have a direct impact for 2021, unlike the other three, Roese said. "It's real, but viability is still many years in the future," he said.
"Last year, quantum computing was still a work in progress, and very rare. Most people had no access to it. This year we have enough quantum systems that are now present...that this year, the broader software development ecosystem will be able to experiment with quantum."
"Quantum should be viewed as an augmentation to conventional compute, not a replacement, when it is viable – they have different mathematics and different functions. It's an 'and', not an 'or'," he added.
"Specialised quantum programming skills can be developed now via simulators or early sandboxes – this will be the first year we get exposure witout having to spend hundreds of millions of dollars," he further predicted.
One of the focuses for quantum would be to work on cryptography updates, he said, as quantum systems would be likely to break the encryption we have today.
Silicon (semiconductors)
Compute is no longer scaling fast enough, so the industry is focusing on specialised compute, accelerators to make certain functions to run at much higher efficiency and performance levels, Roese said, predicting a restructuring of the semiconductor ecosystem to feature core processing that is paired with domain-specific architectures (accelerators) such as GPUs.
Semiconductor companies have been positioning themselves for a new era with relevant acquisitions, Roese said, naming Intel with Habana and Altera; NVIDIA with Mellanox and ARM; and AMD plus Xylinx as three such examples.
"We will need different software frameworks and will also need new integration platforms (future servers)," he said.
"This change in the way compute exists will happen not just technically but at the industry level this year."
Roese said various parts of the Asia Pacific region are likely to be front-runners for the four trends. "I'm very optimistic that while not all of Asia will be first, somewhere in Asia.. is likely to be one of the first movers in one of the trends I described," he said.
“The key piece is the innovation centre of gravity is moving towards Asia, both from a consumerism perspective as well as an innovation perspective,” added Midha.
No comments:
Post a Comment