Data-Center-Class File System
Stellus takes a unique, modern approach to solving unstructured data challenges at the cloud, in the core, and on the edge infrastructure necessary for today’s most successful businesses. Stellus leverages the growth models of the largest hyperscale companies and combines high-throughput key-value stores, high-performance key-value over NVMe Fabrics networking, and algorithmic data placement into a single data-center-class solution. The key-value store approach removes common bottlenecks associated with data access at scale, including resource contention costs and scale and performance limitations of global data caches and global maps. Stellus enables direct access to high-performance memory and NAND, which translates to more than twice the application performance with I/O throughput of 10-GB/s per Data Manager (DM) for both read and write—unlike other systems in which write speeds are a fraction of reads.
At the same time, Stellus provides file access across industry-standard protocols like NFS—all in a scalable, Cloud/Core/Edge ready platform. The Stellus Data Platform works as a complete file system, so users can take full advantage of the latest infrastructure available in Cloud catalogs and Core networks.
High-Performance Scalable Data Platforms
Based on key-value store technology and purpose-built for the world’s unstructured data. The Stellus Data platform is designed for the most demanding data-intensive workloads and applications, such as high-speed microscopy, genomic sequencing, machine learning, cognitive AI, and data streaming at cloud and in core infrastructures.
The Stellus Data Platform is a software-defined and delivered file system that provides hyperscale throughput for digital enterprises. This platform allows businesses to capture, store, and process increasingly large data sets, and it makes data accessible faster and more efficiently than storage infrastructures available today. Stellus helps organizations scale through data challenges by decoupling capacity and performance in a near-linear trajectory.
Key-Value Store Scalability
Modern unstructured workloads require a modern enterprise data strategy. Stellus key-value stores serve data at stellar speeds, dramatically improving performance and reducing latency and strain on CPU, memory, and energy.
Decoupled Performance & Capacity
Decoupling performance & Capacity in a near linear trajectory allows users to provision and scale data store capacity and data store performance independently.
Users can scale system performance and capacity while seeing consistent resource utilization for data placement, consistently low latency as the system scale increases, and consistent processing of data requests.