Data is central to today’s modern enterprise.

But many organizations don’t always know how to use their data, and they struggle to access it, as it is spread across disparate locations (on-premises, in public clouds or at the edge). Many assume they have to take on the arduous task of consolidating all that data in one place before they can even begin to use it.

This doesn’t have to be the case, according to Dell. With new announcements today, the tech giant aims to help organizations embrace data gravity and access data where it is. This, the company says, can accelerate how quickly data scientists can access and use their data — wherever it is — and rapidly build out critical artificial intelligence (AI) and generative AI (genAI) tools.

“We’re all familiar with the continued explosive growth of data,” Martin Glynn, senior director of product management at Dell, said during a press briefing. “Our goal is to help support our customers to get a handle on this data and be able to use it as soon as possible.”

According to Dell research, 92% of business and IT executivess agree their organization needs to shift to an AI-first operating model within the next 12 months to stay competitive.

“The way they’re going to do that is to use their own data as the differentiator,” said Glynn.

He pointed out that the total “global datasphere” will hit more than 175 zettabytes by 2025. Of that amassed data, 90% is unstructured (such as videos, images, documents and audio) and the remaining 10% is structured data stuck in silos.

New PowerScale OneFS software updates — announced today and available later this month — will enable enterprises to wrangle that data and meet increased demand for AI-ready infrastructure and faster performance.

The updates will help organizations prepare, train, fine-tune and inference AI models more quickly, Glynn said. Dell asserts that its new PowerScale all-flash storage systems will provide up to 2X performance increase for streaming reads and writes.

“That performance gain is particularly well-suited for AI environments,” said Glynn.

PowerScale OneFS will extend capabilities into an “open ecosystem,” he said, where enterprises can work with Dell partners with different toolsets in data analytics, AI/ML workflows and with cloud capabilities. This is all built on Dell compute, Dell storage and Dell Validated Design.

“At the heart of this data platform is our ability to connect,” said Glynn.

PowerScale will also offer a new smart scale-out capability to improve compute node performance and enhance GPU use. Dell plans to deliver new all-flash nodes in the first quarter of 2024.

Furthermore, enhancements to Apex  file storage for AWS will offer larger cluster capacity and more OneFS features, according to Dell. Additionally, Dell Apex for File Storage for Microsoft Azure will be globally available in the first half of 2024.

“Storage performance is a critical factor for successful AI and generative AI outcomes,” Arthur Lewis, president of Dell’s infrastructure solutions group, said in a statement. “Customers are relying on us to continually push the boundaries of storage innovation, including removing data access bottlenecks that limit the throughput and scalability of compute-intensive applications.”

Also announced today, Dell PowerScale is expected to be the first Ethernet storage solution validated on Nvidia DGX SuperPOD. Ethernet has the capability to increase flexibility, speed and ease for AI storage, Glynn said.

Through the collaboration, customers can take advantage of a combination of Nvidia DGX systems, Dell PowerScale storage and Nvidia Quantum-2 InfiniBand and Spectrum Ethernet networking. This can help them accelerate AI and genAI initiatives with Dell’s network-attached storage. The capability is expected in the first quarter of 2024.

Dell is also extending capabilities around multicloud to enable genAI business. Dell APEX file storage for AWS and Microsoft Azure — available in the first half of 2024 — will allow organizations to take advantage of AI workflows on data in public clouds or on-premises.

Glynn said this will offer flexibility across multicloud environments, enabling customers to choose from a variety of large language models (LLMs) and use libraries from Hugging Face or MosaicML to retrain models on their proprietary data stored with Dell.

Dell Apex for file storage integration with Databricks and MosaicML is available in AWS today and in Microsoft Azure in the first half of 2024.

Integral to all this is Dell’s planned open modern data lakehouse — announced in conjunction with Starburst in October and expected in the first half of 2024 — which offers a “single point of data access,” Greg Findlen, SVP of product management at Dell, said in the pre-briefing.

The modern data lakehouse is “multicloud by design,” with the ability to store, process, manage and secure data across on-prem, cloud and edge environments, he said. The platform can reduce data movements, allowing organizations to discover, query and process data where it is.

“There’s a huge amount of unstructured data sitting out there. Customers need to be able to process that along with their structured data that has a lot of critical information,” said Findlen. “Our objective is to reduce data movement, helping customers leverage data gravity and have the potential to unlock that data without it having to all be in one place.”

High-performance engines in the planned data lakehouse will be able to query and process data for different workloads. A unified data stack can handle “any workload” from BI to AI to ML, and “any data,” whether structured, semi or unstructured, Findlen said.

Furthermore, the modern data lakehouse “needs to have an open, scale-out architecture,” he said. “A lot of historical or legacy designs have really coupled compute and storage from scaling perspectives. That has created a lot of limitations and burdens on customers.”

Security and governance are critical, too. “You want to ensure that the right people are seeing the right information,” said Findlen. This requires access control, masking, encryption and data localization.

Findlen emphasized that “open is a key part” of the lakehouse. Dell will allow customers to leverage the latest trends in open source while benefiting from automated deployment and ongoing management.

“The objective is to make sure we can help customers discover and integrate and process data across the organization,” said Findlen.

Dell also this week announced an expansion of the Dell Generative AI Solutions portfolio with semiconductor provider — and Nvidia rival — AMD. As part of the partnership:

• Dell PowerEdge XE9680 will support the forthcoming AMD Instinct MI300X accelerator. This enhanced silicon diversity will help organizations access genAI and train their own LLMs.
• Dell Validated Design for Generative AI with AMD: Will support AMD ROCm-powered AI frameworks.