Getty Images

Vast Data sets sights on analytics, AI

Vast Data expands on its primary storage origin story, adding database management options and targeting analytics, all with an eye on AI workloads.

With a new product, Vast Data is merging storage and data management together, narrowing the gap between the two as AI continues its rapid market expansion.

Vast Data's Universal Storage, an all-flash, share-everything, scale-out file system, has become Vast Data Platform, which adds database management and analytics capabilities to its existing storage offering. The new features don't change the price of Vast's pure storage offering, and customers happy with Vast as a storage provider don't have to use the new features. The move comes at a time of exploding interest in AI, which has sparked a rethinking of storage and its role moving forward.

"Some of the most interesting and fastest-growing workloads right now are in the advanced analytics space, especially around AI, gen AI, machine learning and HPC [high-performance computing]," said Dave Pearson, an analyst at IDC.

While they might be the most interesting, they do not present the biggest capacity challenges or make up the majority of use cases for storage, Pearson said.

Expanding beyond storage to tackle AI

Building off human intelligence has advanced society throughout history, and artificial intelligence will amplify this, potentially bringing solutions much quicker, said Max Tegmark, a professor of physics and AI researcher at MIT, at Vast's recent Build Beyond show. Finding new ways to process the data is key to unlocking AI's potential.

"It's the structure of information processing that matters, nothing else," Tegmark said.

Originally, Vast provided an element store for files and objects and a write buffer that increased the efficiency of quad-level cell flash while adding features such as data protection and encryption.

With the Vast Data Platform, the company has added a table and the standard query protocol. The new additions allow SQL queries to be run on file and object metadata, said Renen Hallak, co-founder and CEO at Vast, during the Build Beyond unveiling.

This ability enables unstructured data to be analyzed and processed faster, which is critical for AI workloads given that unstructured data is the basis for all forms of modern deep learning, according to Jeff Denworth, co-founder and chief marketing officer at Vast.

The DataStore is the vendor's high-performance storage, with NAS protocols, network file system protocol for parallel file processing, HPC and support for emerging AI workloads, he said.

The output of AI is transformative, and storage continues to play a central infrastructure role, Denworth said. He added that while Vast Data is categorized as a storage vendor, its focus has been on how to best store data to serve an application. The data stored on AWS, Azure, Google Cloud Platform and Oracle is central to what companies do, but none of the cloud providers is viewed primarily as a storage vendor.

"Storage is at the heart of the system, but the system is more than just storage," Denworth said.

Vast expands, storage market stays the course

Denworth said the transition from Universal Storage to the Vast Data Platform was an expansion of the technology Vast was already offering, which enables layering rather than pivoting.

"Expansion is the easiest way to layer on top of a business, and that also happens with the product [such as this] -- it's an expansion of capabilities and the customer set," Denworth said.

The new data platform is a natural extension of Vast's business, according to Steve McDowell, an analyst and founding partner at NAND Research. Vast is going from large, fast storage to an immediate adjacency of analytics and AI that use fast storage.

Analytics stacks are complex, and moving data closer to the analytics could provide for better performance, McDowell said.

"This simplifies analytics by compressing the stack," he said.

The storage market as a whole won't shift to providing combined data management and storage services, but others will follow Vast in offering products aimed at simplifying analytics, McDowell said. He believes those that follow would most likely be file system vendors such as Weka, Panasas or IBM.

The release of Vast Data Platform is the vendor's attempt to adapt to where it perceives the market is going, IDC's Pearson said. IT is good at storing large amounts of unstructured data and good at using relational databases; combining the two would be the next step.

"Overlaying those two together while avoiding the typical limitations of both is what we are looking for right now in terms of advanced analytics," Pearson said.

In with the new, keep the old

The combination of analytics and storage could change how data is stored, in Denworth's estimation. As AI becomes more prolific, companies will find new needs for data they would normally archive away. Deep learning, for example, improves as more data is exposed to learning models, he said.

The growth of data is the rising tide that is lifting all storage vendors.
Dave PearsonAnalyst, IDC

"I think the classic storage paradigms aren't necessarily appropriate for the changing application paradigms," Denworth said.

Pearson said a company's storage strategy still depends on its use cases. There are still many examples where storage scale, density and archive are the main drivers. Data management is a large element of worldwide data storage, but it isn't driving out other storage workloads. Over the last 20 years, data growth has been the main driving factor in storage, and data management won't be the endgame.

"The growth of data is the rising tide that is lifting all storage vendors," he said. "But finding the niches in storage that deliver real business value will differentiate vendors and increase revenue."

One niche is generative AI, where storage vendors are looking at how they can "hook their wagons" to it, McDowell said. Vast was already there with its performance and file system. Other vendors will focus on delivering data to AI. The Vast Data Platform catalogs metadata for quick retrieval and delivery to analytics. While this is an interesting product, it is a niche, he said. The market for storage and analytics combined is only so big, for now.

Adam Armstrong is a TechTarget Editorial news writer covering file and block storage hardware and private clouds. He previously worked at

Dig Deeper on Storage architecture and strategy

Disaster Recovery
Data Backup
Data Center
and ESG