SHENZHEN, China, May 27, 2025 /PRNewswire/ — AI applications incur significant computing demands, requiring cloud infrastructure and storage to be elastic and scalable. To help enterprises cope with increased workload capacities, Tencent Cloud recently upgraded its GooseFS 2.0 product to support better optimization for various AI scenarios. The upgrades will see greater enhancements for computing, metadata and server-side accelerations, enabling enterprises to gain higher performance and accessibility from their data lakes.
Under the new architecture, GooseFS 2.0 maximizes hardware performance through short-circuit reads and metadata memory acceleration, easily achieving single-link speeds at the GBps level, approaching the limits of hardware resources. It addresses the challenge of hotspot distribution through secondary caching, balancing cache space usage with high QPS/bandwidth capabilities, thereby improving the efficiency of large model distribution.
GooseFS 2.0 also leverages on hierarchical metadata management technology to enable horizontal scaling of file clusters, and achieves millions-level metadata IOPS. Optimizations such as cross-node hard links enable disaster recovery capabilities at the second level, even with metadata scales in the tens of billions.
