Industry executives and experts share their predictions for 2022.  Read them in this 14th annual VMblog.com series exclusive.

Introducing a Year of End-to-End, Virtualized AI

From AI on hybrid clouds to 5G at the telco edge - and even enterprise virtual worlds and virtual reality - you might not realize how many advanced workloads will run virtualized in 2022

By Justin Boitano, vice president and general manager, Enterprise and Edge Computing at NVIDIA

It's hard to believe that 2022 marks nearly two full years since the pace of digital transformation skyrocketed due to the COVID-19 pandemic. As the world began working remotely, IT teams responded by delivering more services and more infrastructure through virtualization and remote access to enterprise infrastructure. With few traditional options available during lockdowns, users embraced new ways of getting things done, accepting and adopting new workflows with a speed that few of us might have ever seen before in our careers.

IT has had to make this shift while simultaneously supporting wider use of AI and accelerated computing in the enterprise. Image recognition, recommender systems, natural language processing and adaptive threat detection are just a few of the AI-powered solutions that have gone mainstream. Enterprises are now deploying AI technology to improve sales, product design, customer service, manufacturing, cybersecurity and a dozen other business aspects.

They say that necessity is the mother of invention. Luckily, advances in computing and software meant many enterprises around the world were well prepared for these shifts, at least from a technology capabilities standpoint. Now, from AI to VR, workloads that once seemed impossible to virtualize and manage at scale are being deployed to advance business transformation across industries.

Of course, innovation-and time-marches on. A number of enabling technologies are key to making 2022 the year of end-to-end, virtualized AI in the enterprise. Here's how they'll each contribute:

• Hybrid clouds power advanced AI on enterprise infrastructure: As 2022 begins, enterprises now have an easy entry path to the era of AI and will be running AI-powered services alongside their traditional applications on their hybrid clouds and core data centers. We reached this moment through a number of key technologies that just became available in 2021. A cornerstone is the NVIDIA AI Enterprise software suite, which is optimized, certified and supported by NVIDIA for VMware vSphere with Tanzu. This AI-ready enterprise platform makes it possible to run advanced AI training and inference workloads on mainstream, NVIDIA-Certified servers, centrally managed by IT.

The joint NVIDIA and VMware offering is the best of both worlds: IT teams can manage AI workloads with familiar vSphere management tools like vCenter. AI practitioners and data scientists can innovate in an agile, cloud native Kubernetes environment based on VMware vSphere with Tanzu and NVIDIA AI software.  Additionally, the recently expanded NVIDIA LaunchPad program helps enterprises around the world test and prototype AI workloads at no charge with NVIDIA AI Enterprise.

• DPUs transform data center AI: 2022 will be the year of software-defined, hardware-accelerated data center networking that supercharges AI performance by offloading infrastructure workloads from CPUs to DPUs. Many AI workloads are distributed and demanding, hungry for CPU cycles, network bandwidth and storage performance. Security best practices and data privacy rules require strong boundaries between users and virtualized workloads, plus robust security on every part of the network.

The availability of NVIDIA BlueField-2 DPUs in 2021 equipped enterprises with additional performance, security and manageability for their data centers. DPUs place a "computer in front of the computer" for each server, delivering separate, secure infrastructure provisioning and acceleration that is isolated from the server's applications. This allows agentless storage virtualization, hardware-offloaded SDN, hyper-precise time synchronization and smart telemetry on both virtualized and bare-metal servers. BlueField DPUs are also a key technology in NVIDIA's Zero-Trust Cybersecurity Platform, announced at GTC 2021. Innovators including Baidu, Palo Alto Networks, Red Hat and VMware are using NVIDIA BlueField DPUs to transform data centers and bring differentiated offerings to market.

As 2022 accelerates the pace of innovation in AI and cloud-native data center architectures, the forthcoming BlueField-3 will deliver further breakthroughs in performance, powering the next wave of secure, accelerated cloud and AI computing applications.

• AI moves to the telco edge: 5G will open new opportunities for edge computing. Key benefits will include network slicing that allows customers to assign dedicated bandwidth to specific applications, ultra-low latency in non-wired environments and improved security and user/workload isolation.

AI-on-5G will unlock new edge AI use cases. Much of new AI inference will happen close to where the data is generated and where insight can be applied. These include "Industry 4.0" use cases such as plant automation, factory robots, monitoring and inspection; automotive systems like toll road and vehicle telemetry applications; as well as smart spaces in retail, cities and supply chain applications.

This year, expect to see more maturity in the AI-on-5G solutions in the market that will enable enterprises to take advantage of private 5G. Look for full-stack platforms that provide the performance, management and scale of these environments.

There will be multi-tenancy across industries and every business function. Telco edge will run intelligent video analytics next to extended reality (XR) apps in VMs and containers, while AI in sales/marketing for personalized recommendations might run workloads in the same cluster next to virtual assistants or chatbots, all separated and managed by familiar tools such as VMware vSphere with Tanzu and NVIDIA Fleet Command.

• Remote collaboration software and hybrid workspaces support seamless transitions between office and home: From creators and students to engineers, federal workers and data scientists, people around the world are now logging in from outside of traditional workplaces and many will continue to do so. Yet keeping up with complex workloads like interactive graphics, data analytics, machine learning and AI requires the powerful performance people relied on in the office, lab and classroom.

In 2022, enterprises will enter a new era of 3D design, collaboration and simulation through software platforms like NVIDIA Omniverse Enterprise running on NVIDIA-Certified Systems, NVIDIA vGPU virtualization software and the Project Maxine SDK for video conferencing. These solutions ensure people can tackle critical day-to-day tasks and compute-heavy workloads wherever they might be working and open new opportunities for creating in virtual worlds and benefiting from digital twins.

• AR/VR modernizes enterprise design and collaboration: CloudXR, NVIDIA's streaming AR and VR SDK, is now integrated into VMware's Workspace ONE XR Hub providing production-level XR streaming. Users can now quickly and more securely access complex virtual and augmented environments, scenes and simulations running on powerful remote workstations using an all-in-one headset. Enterprises can easily deploy XR-capable virtual machines accelerated by NVIDIA virtual GPU software. Streaming XR from remote, high-powered graphics servers changes the distribution of high-fidelity, photorealistic XR environments and capabilities, essentially democratizing XR for anyone with a mobile XR device.

As 2022 nears, it's clear that virtualized enterprise computing is evolving quickly. IT teams are moving to support new applications that help businesses run new workloads in new ways on accelerated enterprise infrastructure. Users will benefit from AI, accelerated infrastructure, 5G and VR/AR as they work and collaborate in ways that were hard to imagine just two years ago. The new year will surely bring more surprises, but it's a safe bet that companies will embrace more data center-scale computing with full-stack solutions as they prepare for what lies ahead.

##

ABOUT THE AUTHOR

Justin Boitano leads the enterprise accelerated data center business at NVIDIA. Previously, he was vice president of marketing and business development at Frame, a multi-cloud app delivery service (acquired by Nutanix) and was general manager of NVIDIA's enterprise virtualization and cloud business. Justin draws on his 19+ years of business leadership, a B.S. in computer science and an MBA to identify high-value opportunities to help enterprises unlock previously unrealized productivity gains with NVIDIA GPU acceleration from AI to VDI.