VMware Inc.

04/08/2021 | Press release | Archived content

New Strategies to Fast-Track AI in the Enterprise

Cloud
/ April 8, 2021 / Blakely Thomas-Aguilar

AI is ready for every enterprise looking to transform its business. Use cases abound across diverse vertical markets where AI is helping financial service providers improve customer relationships, allowing manufacturers to streamline supply chains, and giving patients better outcomes from their doctors.

Most organizations know they need to invest in AI to innovate quickly, stay competitive, and secure their future. But leaders struggle to find the strategy and platform that enables success. That's particularly true for AI applications.

3 Challenges Stand in the Way of AI Deployment

Unlike traditional enterprise applications, AI apps are a recent development. These apps are anchored in rapidly evolving and bleeding-edge code. Unfortunately, this means AI apps also lack proven approaches that meet the rigors of the enterprise in a scaled production setting.

Simply put, AI apps can be:

  1. Complex and slow to deploy.
  2. Hard to scale cost-effectively.
  3. Brittle to manage.

Challenge #1: Risk & Complexity

It's difficult to pull together an end-to-end AI solution from disparate parts and integrate it within the existing infrastructure. That's why just 53% of projects make it from pilot to production, according to Gartner's 'AI in Organizations 2020 Survey.' The complexity of AI solution integration with existing infrastructure is among the top three barriers to implementation.[1]

[1] Gartner. 'P-19019 AI in Organisations.' Claudia Ramos, Erick Brethenoux, 2020

Challenge #2: High Performance

High performance is critical to AI, machine learning (ML), and data analytics workloads. Data engineers need to crunch through vast amounts of data quickly. Data scientists need to develop and train their models quickly for fast time to deployment. They also need to be able to scale up performance within a single server or scale outperformance through multiple servers, depending on the workload.

Challenge #3: Scale

Going from proof of concepts to enterprise deployments needs effective scaling. This means making efficient use of precious GPU resources, as well as enabling manageability and availability. The costs of provisioning siloed bare metal AI infrastructure or public cloud infrastructure can also be challenging.

To pull insights from data, enterprises need AI-ready infrastructure that delivers substantial compute power and high-performance access to data, along with the right tools and algorithms for their teams. And now more than ever, leaders need to transform their businesses, while delivering a positive return on investment and scalable infrastructure.

For enterprises to stay competitive, it's no longer a question of if they should modernize their infrastructure for AI. It's a question of how quickly they can evolve their infrastructure to do so while managing risk.

Krish Prasad, SVP and general manager, VMware Cloud Platform Business

Strategies to Fast-track AI in the Enterprise

Fortunately, infrastructure innovation is catching up to the needs and demands of today's AI apps. AI workloads require a new platform that's optimized for the enterprise and supports the full lifecycle of implementation:

  • Experimentation.
  • Prototyping.
  • Model training at scale.
  • Deployment in production.

This new platform must offer a unified, scalable solution that solves the three key challenges and withstand the rigors of an enterprise setting.

1. To combat risk, businesses need an end-to-end solution that supports existing infrastructure.

Organizations need to streamline deployment and management of AI, machine learning, and data analytics workloads. An end-to-end solution helps businesses avoid AI silos and simplify management. One of the most recent innovations leverages virtualization to fold AI deployments into existing enterprise infrastructure. Ultimately, this will help accelerate AI adoption in the enterprise.

2. To move faster, businesses need solutions that deliver world-class performance.

Compute-heavy requirements for AI workloads must be satisfied with a new approach. One is example is near bare-metal performance for virtualized GPU-powered workloads, such as deep learning training and inference. Another is scaling performance up or out by leveraging GPUs in the same server or multiple servers. The solution must also enable faster time to market by running projects and training models quickly.

3. Organizations need to scale without compromise.

With data engineers and scientists needing faster, more reliable resources, solutions must flex based on demand. Solutions also need to enable high availability and simplify infrastructure maintenance, such as consolidation, expansion, or upgrades. Finally, capabilities should also include simpler management at scale, automatically placing workloads across AI infrastructure for optimal resource consumption.

'Overall, it's clear that AI will transform every industry, and we're just at the beginning of integrating this insight into enterprise applications,' wrote technology industry analyst Maribel Lopez on Forbes. 'Companies need new technology solutions that can speed the pace of deployment, help IT leaders add new insights into applications, and create new products. It's good to see new products and collaborations coming to market to support this.'

Innovation Spotlight: NVIDIA and VMware

In early 2021, NVIDIA and VMware delivered an end-to-end enterprise platform optimized for the most powerful AI apps. This platform will:

  • Accelerate the speed at which developers can build AI for their business.
  • Enable organizations to scale AI workloads on existing VMware infrastructure.
  • Deliver enterprise-class manageability.

'With VMware and NVIDIA's partnership, organizations can finally extend the benefits of a common enterprise infrastructure for their AI initiatives,' said Ashish Nadkarni, group vice president, IDC. 'A single platform for consolidating AI and enterprise workloads provides better security, manageability, and resiliency for AI workloads. Customers can now accelerate their AI initiatives with confidence, knowing that such initiatives will scale well.'

Learn more about the NVIDIA-VMware partnership and joint collaboration.

Innovating for AI's Future in the Enterprise

Leading technology innovators are partnering with customers to turn AI's promised benefits into real-world value. Now, enterprises can begin to leverage these emerging solutions to fast-track AI and deploy modern, AI-powered apps at scale. Ultimately, these innovations have the potential to bolster the way businesses go to market, support customers, and empower employees.

Dive deeper:

Our Stories. Your Inbox.

Get our most popular articles, videos, and more-recommended just for you.

Categories: Trending