SandStone CNFS

SandStone CNFS is a high-performance file storage for AI training, machine learning, big data analysis and other scenarios. Through client distributed cache, high-performance independent metadata architecture and other technologies, it gives full play to the performance of near-line cache, reduces data access latency, and uses high-performance storage power to enable and improve the processing efficiency of computing power business.

Download Apply for trial

Product Advantages

Strong features, comprehensive

High concurrency and low latency, speed up the efficiency of data analysis and decision-making

By providing high concurrency and low latency data access, it empowers and improves the efficiency of data analysis and decision-making ability, so that enterprises can stay ahead in the fierce market competition and quickly adapt to market changes.

Massive high reliability, simple operation and data management

Through highly reliable storage technology, it ensures the long-term preservation and integrity of data and reduces the risk of data loss. The system provides resource optimization suggestions according to the usage patterns and trend changes, helping users make more effective use of storage resources and improving the efficiency of data management and operation and maintenance.

Expand capacity on demand to reduce initial investment costs

By horizontally scaling storage resources on demand with the growth of data volume, resources are allocated according to the current needs and budget of the business to avoid over-investment.

Product Overview

SandStone CNFS makes AI training and big data analysis more efficient

Product Features

High performance, high reliability, high expansibility, strong compatibility

High concurrency, high throughput, low latency data access

By evenly distributing data to multiple server nodes, and each server node can provide business access, the hardware performance is fully utilized to meet the high concurrency and high bandwidth data access requirements of big data analysis platform. The client-side cache provides access capabilities close to local SSD performance, and combines multi-level cache acceleration technology to reduce data access latency and improve the efficiency of business processing such as AI jobs.

Massive, scalable and reliable storage

The data and metadata engine adopts advanced distributed architecture to support horizontal scalability of performance and capacity synchronously, and easily supports the challenge of hundreds of billions of small files and EB-scale capacity storage to meet the needs of massive data storage. It supports strong consistent data writing, data copy /EC, disaster recovery, backup and other data protection mechanisms, is compatible with multi-cloud ecology, and provides highly reliable data storage.

Multi-protocol interface compatibility

It provides POSIX, HDFS, S3/OSS storage interfaces, as the unified storage of big data, cloud native data warehouse, AI job and other business platforms, and meets the different storage interface access requirements of each business system. It provides Kubernetes CSI Driver support to seamlessly access cloud native ecological services.

Typical application scenarios

See if SandStone CNFS is right for your scenario?

Big Data analytics platform

Distributed storage provides key support for big data analysis platform. Through its horizontal scalability and high performance characteristics, it can efficiently process massive data to meet the challenge of petabyte data volume. Flexible data access supports real-time and batch analysis needs, accelerates business insight and decision making processes, and ensures a rapid response to real-time data analysis.

AI training/machine Learning platforms

Support the growing demand for datasets and models in AI training and machine learning through scalability and resilience Through parallel processing and fast data access, the model training process is accelerated, so as to shorten the development cycle and improve the efficiency of research. Through the high throughput and low latency data transmission capability, it ensures the smooth execution of large-scale parallel computing tasks.

Cloud native data warehouse/data lake

Through efficient metadata management, the data index and query of cloud native data warehouse and data lake business are more efficient, and the process of data retrieval and analysis is accelerated. Through automated data management functions, such as data lifecycle management, automatic hierarchical storage and other features, the usage of storage resources is further optimized and operational costs are reduced.

Microservice Container Platform

It supports a unified namespace and global access view, allows different containers and applications to share and access the same data set, promotes data consistency and collaboration between teams, and simplifies data management of microservice applications. Through the high availability and data redundancy mechanism, the risk of service interruption caused by storage failure is reduced, and the stability and reliability of the platform are enhanced.