Lightbits
Labs published a
reference
architecture (RA) with
AMD, for the industry's first scalable, highly
available, disaggregated, software-defined storage solution that simplifies and
accelerates Kubernetes infrastructure deployments. This game-changing solution
delivers unmatched price/performance value, enabling a seamless path to
application containerization and modernization, while making it easier to
manage and scale infrastructure for high-performance and AI workloads.
The RA is designed to leverage the combined power of Ceph
and Lightbits storage, with AMD hardware, to ensure an efficient and scalable data
platform optimized for general-purpose applications, as well as demanding AI and database applications in
Kubernetes environments. A key
advantage is the seamless integration of Lightbits block storage into existing
Kubernetes deployments, without requiring significant re-architecting, which
enables a smooth upgrade path to software-defined, high-performance block storage.
"Organizations
with application containerization or modernization initiatives are sensitive to
cost-efficiencies while needing scalable, highly performant storage
solutions," said Kam
Eshghi,
Co-Founder and Chief Strategy Officer at Lightbits Labs. "Our reference
architecture with AMD provides a fast, simple on-ramp to software-defined
infrastructure for Kubernetes deployments, delivering superior performance for
latency-sensitive workloads, without the need for additional costly changes to
network architecture."
Designed natively for NVMe over TCP, Lightbits software-defined storage redefines performance and
efficiency for stateful applications in Kubernetes. The storage software scales
to hundreds of petabytes and achieves up to 75M IOPS per cluster with
consistent sub-millisecond tail latency, making it ideal for
performance-intensive applications. Lightbits software running on AMD EPYCTM
CPU-based servers yields a cost-efficient solution that radically simplifies
storage management and accelerates business outcomes, making it easier to
provision, manage, and scale high-performance, persistent storage for Virtual
Machines (VMs). The system architecture can be easily scaled to support many
Kubernetes worker nodes while maintaining linear performance.
Additional benefits of this solution include:
- Support for mixed general-purpose
and high-performance workloads on a single storage solution
- Operational simplicity with the
ability to provision and scale resources via the CSI plugin
- Sustainability through more
efficient utilization of hardware resources, smaller data center
footprint, and lower costs
- Versatility to support both
containerized and virtualized (using KubeVirt) applications
"Modern cloud-native workloads demand performance,
efficiency, and seamless scalability at every layer of the stack," said Derek Dicker, corporate vice president,
Enterprise and HPC Business Group, AMD. "By pairing AMD EPYC processors with
Lightbits' software-defined storage, customers gain exceptional compute and
ultra-low-latency storage that can accelerate even the most data-intensive
Kubernetes pipelines."
To get started building a data platform for high-performance block
workloads in Kubernetes environments, download
the reference architecture available on the Lightbits Labs website.