CTERA, a leader in intelligent data management, is introducing CTERA Fusion Direct, a transformative federated data architecture that eliminates the historical trade-off between enterprise file systems and object storage.
According to the company, by unifying these two domains into a single, high-performance data fabric, CTERA provides the foundational infrastructure required for the scale and speed of today’s AI workloads.
The Fusion family, consisting of CTERA Fusion Gateway and now CTERA Fusion Direct, focus on creating a unified fabric between NAS and Object storage.
CTERA Fusion Direct is built on a fundamentally new architecture that enables files and objects to coexist natively within a single, federated global namespace. Data can be written as files and read as objects, or written as objects and read as files, with full bidirectional read and write capabilities. There is no file-to-object conversion bottleneck, no proprietary chunking scheme, and no translation gateway mediating object access, the company said.
Enterprise applications and humans can access data over SMB and NFS exactly as before. Simultaneously, AI training clusters, HPC environments, and cloud-native services can access the same data over S3 and S3 over RDMA for wire-speed throughput.
Existing object storage buckets can be attached directly to the data fabric, enabling seamless access to contained data across edge locations and multi-cloud environments without migration, duplication, or modification. Organizations can instantly present object data as files on a global scale, activating distributed datasets across geographies while reducing infrastructure complexity, the company said.
Backed by U.S. Patent 12,007,952, CTERA Fusion Direct is engineered to leverage CTERA's core technology stack, maximizing both human collaboration and machine-scale data consumption:
- Native Zero-Copy Access: Data written through the CTERA Intelligent Data Platform is immediately available as standard S3 objects. Conversely, existing S3 buckets can be attached as-is, making their contents instantly accessible as files. No duplication, no latency, and no storage overhead.
- High-Speed File Streaming: High-resolution media and massive datasets can be streamed directly from object storage to file-based applications, eliminating the need for time-consuming local downloads.
- S3 over RDMA and GPU-Direct Access: By exposing native objects, CTERA Fusion Direct supports high-performance standards such as S3 over RDMA, allowing AI clusters and GPUs to read and write data at maximum bandwidth.
- Future-Proof Data Sovereignty: Data remains in standard S3 buckets, free from proprietary encapsulation or translation layer. This ensures organizations maintain full control over their information assets across any cloud or on-premises platform.
“The primary roadblock to enterprise AI has not been a lack of data, but the inability to use it effectively. The historic divide between data for human collaboration and data for machine analytics created friction and complexity that slowed companies down," said Oded Nagel, CEO of CTERA. "We've collapsed those silos. This fundamental shift in architecture unlocks the full potential of a customer's data, providing the unified data platform necessary to not just compete, but to lead in an economy increasingly driven by machine learning.”
CTERA Fusion Direct is now available as a core component of the CTERA Intelligent Data Platform.
For more information about this news, visit www.ctera.com.