Scality and WEKA have introduced a joint solution that combines WEKA’s NeuralMesh with Scality RING to support enterprise data storage requirements. The integration is designed to deliver high-performance storage alongside scalable object-based capacity for AI and high-performance computing (HPC) workloads.
At the centre of the solution is Scality’s lightweight object connector for NeuralMesh, which enables integration between the two platforms. Compared with traditional S3 interfaces, Scality reports that its object connector achieved up to 10x higher performance on similar hardware in testing, along with up to 20% lower infrastructure costs. The aim is to allow organisations to extend AI data pipelines while managing infrastructure and operational overhead.
The solution architecture keeps active and new data on WEKA’s flash-based NeuralMesh tier, while transitioning older or less frequently accessed data to Scality’s object storage. This approach is intended to reduce reliance on all-flash deployments while maintaining performance for active workloads and long-term storage for larger datasets.
NeuralMesh is positioned to support GPU utilisation, accelerate time to first token, and support AI pipelines with high-speed data access.
Scality RING is designed to scale to exabyte capacity and provide durability of up to 14 nines. The object connector provides a tier for scalable storage with reported cost advantages compared with traditional object storage approaches.
The integration is validated for interoperability and does not require additional engineering changes to implement. Data tiering occurs between the flash-based performance layer and the object-based capacity layer.
The connector is presented as an alternative to community-driven object storage systems, offering enterprise support alongside performance and reliability.
Through this collaboration, the combined solution brings together high-performance storage and scalable object capacity to support AI and HPC environments while aiming to reduce infrastructure costs and simplify management.
