During the 56th edition of the IT Press Tour in California, VMblog had the opportunity to meet with Brock Mowry, CTO and VP of Products at Tintri, which is part of DDN. Mowry provided an insightful overview of Tintri's VMstore platform and how it is evolving to address enterprise needs around virtualization, databases, containers and AI.

VMs are Legacy

While virtualization may now be considered a "legacy" term by some in the industry, the VMstore was traditionally focused on virtualization environments. However, Tintri supports multiple hypervisors with deep integrations, including VMware, Citrix Xen and Microsoft Hyper-V. But one of the company's key differentiators is that Tintri understands workloads at the individual virtual machine level.

"We focus on the individual virtual machine," explained Mowry. "So we're able to provide performance, we're able to provide data protection, observability, a lot of other features that are focused at the individual virtual machine level, not at the larger LUN or export level in traditional storage."

AI has Exploded

While the need for AI is recognized by enterprises, it is important to keep in mind that the process of adopting AI across the organization can be complex. It involves aligning AI strategies with business goals, ensuring data availability and quality, and addressing organizational and cultural challenges. Enterprise-wide adoption requires careful planning, collaboration, and change management efforts.

Tintri has also incorporated a form of artificial intelligence within the VMstore to manage performance, tied to its auto QoS feature which is native within the file system. This AI helps prioritize the workloads that the system was designed for and can even penalize unknown workloads to ensure performance for understood VMs.

Is Kubernetes Taking Over?

Beyond hypervisors, Tintri has expanded its support to SQL databases in a unique way by managing at the individual database level. And with the growing adoption of containers and Kubernetes, Tintri is well established to address this growing space as well.

AI platforms often run in containers, and there is a growing trend of running VMs within containers. This adoption of new technologies, such as containers and container orchestration platforms like Kubernetes, may require organizations to invest in learning and implementing these technologies effectively.

"Kubernetes, we're seeing this really starting to take over. It's matured quite a bit," said Mowry. "We've built a CSI to be able to work with the most common Kubernetes platforms out there. We've also built some integration into the VMware Tanzu."

Tintri's CSI driver works with generic Kubernetes to support a broad set of Kubernetes platforms. By talking to the Kubernetes API, Tintri gains a deeper understanding of the environment and data. The VMstore UI pulls out pods, deployments, nodes, persistent volumes and persistent volume claims so admins can see how data is being consumed. They can drill into individual persistent volume performance, manage QoS, and snapshot and replicate data.

Bringing it All Together

When it comes to AI, Tintri is working with key partners on a smaller scale, with experimental AI/ML projects that are built on top of virtualization. The VMstore's ability to rapidly clone and replicate large file sets is obviously seen as another key advantage here.

"We have some unique abilities around being able to clone and replicate large file sets. And we do see that as a potential advantage for some of these smaller operational models, people that have created an LLM and to distribute it within the enterprise, as well as keep that data private within the enterprise," noted Mowry.

The VMstore manages its file system at a pointer-based level, enabling extremely fast data transformation and duplication without unnecessary capacity consumption. Tintri's QoS capabilities also extend to the individual persistent volume level within Kubernetes, ensuring fair share performance for all projects.

Overall, the versatility of the Tintri VMstore to handle multiple hypervisors, databases, containers and AI use cases, all with intelligent QoS and data management capabilities, provides strong investment protection for enterprises. As Mowry summarized, "You can buy into this. You can refactor from a virtual machine into a microservice or a containerized architecture. You don't need to trade or change your underlying storage investment."

For IT professionals looking for a flexible storage platform that can adapt to changing virtualization, database, container and AI needs, Tintri's VMstore is definitely worth exploring further. Its ability to deeply understand and manage workloads at a granular level, coupled with advanced data services and QoS, make it a compelling choice in today's dynamic IT environments.

##