DDN Redefines AI and High-Performing Computing at Scale with Google Cloud Managed Lustre Innovations

Google Cloud Managed Lustre, powered by DDN EXAScaler technology, is now generally available, delivering a high-performance parallel file system designed to accelerate AI and high-performance computing (HPC) workloads. This fully managed service removes the complexity of infrastructure management while providing the scalability and performance needed for demanding data-intensive applications.

Built on DDN’s industry-leading EXAScaler platform, Google Cloud Managed Lustre offers persistent storage with performance that scales with capacity, up to multiple petabytes. Its architecture is optimized for AI/ML and HPC workloads, maximizing GPU utilization and reducing compute idle time by delivering high-throughput, low-latency data access. This accelerates model training, enables faster insights, improves accuracy, and supports complex AI projects by preventing storage bottlenecks as data grows.

The service enhances AI inference workflows by serving as a shared key-value (KV) cache, addressing challenges posed by large context windows in large language models. By offloading KV cache storage to Managed Lustre, organizations can reduce latency sensitivity and improve responsiveness during model interactions, lowering the total cost of ownership for AI inferencing.

Google Cloud and DDN’s collaboration integrates decades of Lustre expertise with Google Cloud’s global infrastructure and AI ecosystem, creating a foundational capability that helps enterprises solve complex challenges in AI and HPC. The solution is designed for seamless scalability, cost efficiency, and uncompromising speed at scale, meeting the demands of modern AI infrastructure that combines accelerated computing with high-performance storage.

As a first-party, deeply integrated service in Google Cloud, Managed Lustre enables organizations to accelerate innovation without managing complex storage infrastructure, bringing the power of the world’s most scalable parallel file system natively into the Google Cloud environment.