● Covered by 1 source · 1 reportMedium impact1 positive
Google Cloud Managed Lustre Introduces Dynamic Tier for Lower Cost and AI/HPC Workloads
Key points
- Dynamic Tier costs 6 cents/GB*month.
- Provides sub-ms latency for hot data.
- Throughput scales linearly up to 80 PB.
- Supports AI/HPC workloads with single namespace.
New Dynamic Tier for Managed Lustre
Google Cloud has introduced a Dynamic Tier for its Managed Lustre service. This new tier is priced at 6 cents/GB*month and is designed to make high-performance parallel filesystems more accessible and cost-effective for a wider range of use cases, particularly in AI and High-Performance Computing (HPC).
Addressing Historical Challenges
Historically, parallel filesystems have been expensive and specialized, leading to fragmented data storage and separate development environments for AI developers. This often resulted in manual data staging, copying, and managing disjointed namespaces, increasing operational overhead. The Dynamic Tier aims to consolidate these processes into a single, unified environment.
Key Features and Performance
The Dynamic Tier provides sub-millisecond latency for hot data, allowing all data to reside within a single namespace. Throughput scales linearly with capacity, supporting up to 80 PB, while maintaining stable sub-millisecond latency for hot data across tens of thousands of clients. Pricing is a single flat fee, eliminating separate charges for disk media, data movement, or metadata IOPS.
The High-Performance Cache (SSD) offers sub-millisecond read latencies, while the Capacity Pool (HDD) utilizes Google Cloud Hyperdisk throughput, with average read latencies of 10 to 30 milliseconds.
Recommended Workloads
The Dynamic Tier is suitable for workloads such as multi-epoch training and training with optimized fetch sizes, where hot data is promoted to SSD after initial runs. It also supports write-heavy checkpointing, with bursty writes landing directly in the High Performance Cache and older checkpoints transparently demoted to HDD. Rapid checkpoint restores benefit from new checkpoints being written to the High Performance Cache. Additionally, it improves interactive responsiveness for developers performing tasks like git cloning and compiling libraries.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Reporting from
Google Cloud Managed Lustre now offers a Dynamic Tier priced at 6 cents/GB*month, allowing users to store all data in a single namespace. This aims to reduce costs and simplify data management for AI and HPC workloads by providing sub-millisecond latency for hot data and linear throughput scaling.