10x More Powerful Cloud Native Storage, Fully Open Source, and Ready to Transform

Persistent storage has long been one of the hardest problems in Kubernetes.

Teams successfully modernize applications with containers and orchestration. Then performance problems appear when workloads need reliable, high speed storage. Databases slow down. Latency increases. Infrastructure teams over provision hardware just to compensate for storage limits. And when organizations run both containers and virtual machines, many storage platforms force them to maintain separate infrastructure stacks.

These challenges are not just technical frustrations. They translate directly into slower applications, delayed innovation, and higher infrastructure costs.

The Longhorn V2 Data Engine was built to solve this problem. It introduces a new architecture designed for modern Kubernetes workloads while maintaining the simplicity and openness that made Longhorn widely adopted.

Important note. The V2 Data Engine is currently in Preview and should be evaluated in test environments before production use.

Why Kubernetes Storage Needed a New Architecture

The original Longhorn V1 engine delivered resilient distributed block storage that was simple to deploy and operate. It allowed organizations to run stateful workloads in Kubernetes without adopting complex proprietary storage systems.

However, the iSCSI based data path introduced kernel level overhead that limited performance for the most demanding workloads. High transaction databases, analytics pipelines, AI pipelines, and virtualization workloads require extremely low latency and high throughput.

Organizations that needed this level of performance often turned to proprietary platforms with significant licensing costs and operational complexity.

The Longhorn V2 engine addresses that gap.

A New Storage Engine Designed for Performance

The V2 Data Engine introduces a redesigned data path built on the Storage Performance Development Kit (SPDK). SPDK allows storage operations to run directly in user space rather than passing through the Linux kernel storage stack.

This architecture eliminates the context switching overhead that slows traditional storage pipelines and allows storage I/O to run on dedicated CPU cores using poll mode drivers.

In practical terms this unlocks dramatically higher throughput and far lower latency.

Early benchmarking results show major performance improvements compared to the V1 engine. Tests have demonstrated up to 2 to 4 times faster write performance and up to 2 to 3 times faster random read performance. With additional optimizations such as the UBLK frontend and multi queue support, the total improvement over the previous engine can reach up to 10 times the performance of V1 in certain scenarios.

Full benchmark details are available in GitHub.

The architecture and implementation details for the V2 engine can be explored in the official documentation.

Because the platform remains fully open source, organizations gain enterprise grade storage performance without licensing lock in or proprietary dependencies.

Storage That Works Across Containers and Virtual Machines

One of the most powerful aspects of Longhorn V2 is how it works together with SUSE Virtualization.

SUSE Virtualization, powered by Harvester, allows organizations to run virtual machines and containers on a unified Kubernetes based platform. Longhorn provides the distributed storage foundation for this infrastructure.

With the V2 engine enabled, the same high performance storage system can support both container workloads and virtual machines.

V2 volumes can serve as data disks for virtual machines and support operations such as live migration. This allows infrastructure teams to consolidate workloads onto a single platform instead of operating separate virtualization and container environments.

Harvester documentation for enabling the V2 engine can be found here.

The business impact is significant. Organizations reduce infrastructure sprawl, simplify operations, and unlock higher performance storage for both application architectures.

Why This Matters for Platform Teams

Modern platform teams are under pressure from multiple directions. They must support traditional virtual machine workloads while enabling cloud native development. At the same time they must control infrastructure costs and avoid vendor lock in.

Longhorn V2 moves open source storage into a new performance tier. It enables high performance stateful workloads, AI pipelines, and virtualized applications to run on Kubernetes while preserving the operational simplicity that made Longhorn successful.

When combined with SUSE Virtualization, organizations gain a unified platform for containers, virtual machines, and modern application infrastructure.

See Longhorn V2 in Action at KubeCon

The Longhorn community and SUSE engineers will be demonstrating the V2 Data Engine at KubeCon EU.

Visit the SUSE booth to see how open source storage can deliver the performance required for modern workloads while remaining simple to deploy and operate.

If you want to explore the technology yourself, start with the V2 engine documentation and performance benchmarks above and try the Preview release in your own environment.

If your team is evaluating storage options for Kubernetes or planning a virtualization consolidation, SUSE Consulting can help you design and validate the right architecture. Reach out to your SUSE account team to start the conversation.

Open infrastructure is moving quickly. High performance open storage is a critical piece of that future.

We look forward to building it with the community.

SUSE Storage (Longhorn) is a CNCF Incubating project, fully open-source under the Apache 2.0 license. Learn more at longhorn.io and harvesterhci.io.

For the latest updates, visit suse.com/kubecon or connect with a SUSE expert to explore what’s possible for your organization.