Vast partners with Google Cloud to provide customers with AI tools that can be used on premises or in the cloud. It also puts Vast in another place where AI workloads are running.
Vast Data extended its global namespace capabilities across Google Cloud, providing customers with a view of their data wherever it resides and adding its AI data services to the platform.
The Vast Data Platform, which includes storage, data management and analytics, is now available on Google Cloud Platform (GCP), the vendor said at Google Cloud Next '24. Doing so enables customers to expand Vast DataSpace into GCP and gives them more flexibility in hybrid environments as well as access to Vast tools features in a public cloud.
Vast dealt with extreme volumes of data long before it became important for emerging AI-based applications, according to Merv Adrian, an independent analyst at IT Market Strategy. Newer use cases, including generative AI, make vendors that know how to handle large scale data more attractive to customers headed in that direction.
"The size of the corpus that you reason over is increasingly important and allows you to tackle a lot more difficult problems," he said.
As the boundaries of AI move outward into larger systems, customers are going to continually combine on-premises and cloud data and will look for partners such as Vast to help get them there, Adrian said.
Expanding to Google
Vast contends that its disaggregated shared everything architecture -- where all parts, including storage media and CPUs, are separated and shared for independent clustering of compute and storage -- pairs well with GCP's flexible configurations. DataSpace aims to give users transparency over data on premises and in the cloud. It can be used in hybrid, multi-cloud or single-cloud environments.
Initially, the Vast-GCP combination will most likely support hybrid environments, given Vast customers tend to be on premises, Adrian said.
"The ability to use [Vast] in both places is very powerful," he said, pointing to a combination of security and performance the platform brings to on-premises data storage and management plus resources such as access to publicly available data that Google provides.
Expanding its platform to a public cloud provider isn't new for Vast, according to Steve McDowell, founder and analyst at NAND Research, as it formed a similar partnership with AWS last year. He said the timing of these cloud partnerships as well as a partnership to deliver a new AI architecture with Nvidia is no accident.
"Vast is pushing in two directions," McDowell said. "They've always been an on-premises solution, but much of AI is happening on the cloud."
Vast's vision has always been on software as evidenced by its move to decouple its software from its hardware starting in 2019 to provide more flexibility for deployment and use, he said. To compete with vendors such as Weka, a vendor that also offers high performance data services in the cloud, Vast is making its platform accessible in the cloud as well as supporting customers on premises.
Vast still focused on AI
The size of the corpus that you reason over is increasingly important and allows you to tackle a lot more difficult problems.
Merv Adrian Independent analyst, IT Market Strategy
The Vast-GCP partnership might not be a unique offering, given Vast's partnership with AWS, Adrian said. But it could be a win-win for Vast and Google Cloud customers, he said, as it connects Google's AI tools with Vast's experience managing large swaths of data.
"[AI] is the coin of the realm," Adrian said. "Now you have to be able to do it."
In the past, AI cloud storage services have tended to come from parallel file vendors such as Weka or Lustre, according to McDowell. Vast is moving in that direction, partnering with cloud-based GPU providers such as CoreWeave and Lambda. With GCP, Vast is expanding the cloud options while providing an on-premises component too.
"Vast is saying, 'We can do the same thing that Weka can do, but we also have this powerful on-prem story,'" he said.
Adam Armstrong is a TechTarget Editorial news writer covering file and block storage hardware, and private clouds. He previously worked at StorageReview.com.
