Overview

The distributed file storage system HAOSTOR-H100 is a fully distributed storage product that can be scaled horizontally on a large scale. It organizes local storage resources through storage system software and dedicated hardware, constructs a fully distributed storage pool, provides file system services to upper level applications, and meets the IOPS, bandwidth, and massive expansion needs of structured, unstructured, and semi-structured data access; Providing enterprise level data service features such as load balancing and dynamic grading, helping enterprises easily cope with the flexible and reliable access needs of data in the face of rapid business changes. It has advantages such as large capacity, high performance, low cost, and easy scalability, providing cost-effective storage solutions for smart cities, big data applications, and secure production.

Characteristics

Hardware acceleration: Using FPGA based hardware acceleration to process massive amounts of data at line speed, supporting mainstream data compression encoding and decoding technologies; Through QSFP28/SFP28 optical modules, it can be directly connected to servers or used for device cascading.

Open integration: It can provide a complete POSIX interface and NFS/CIFS protocol, seamlessly integrate existing systems and application software, and achieve full lifecycle management of data.

Super scalability in capacity and performance:

1. The file system provides 300PB level storage capacity, supports billion level file count, reduces system management and maintenance complexity, and eliminates data silos caused by multiple namespaces.

2. Smooth expansion from 3 nodes to 1000 nodes, high availability management of metadata in the Raft triple replica mode, with main data using multi-point to multi-point direct concurrent IO, maximizing the utilization of IOPS/throughput bandwidth on the network and all node disks. Capacity and performance grow linearly with node expansion.

Space quota management:

1. Resource pools support multi-level inter pool data isolation policy configuration, creating secure and isolated storage pools. Improve data security and minimize losses.

2. Support multiple quota management methods, including sequential allocation, average allocation, user specified allocation, etc., to provide customers with easy-to-use, flexible space access, and meet the needs of different customers.

High reliability data protection: N+M data reliability and availability protection mechanism, which can support data protection for M disks or nodes simultaneously failing. Support snapshot and snapshot file system read and write, as well as remote disaster recovery, to improve data storage efficiency.

Simplified resource management: Software platform automates deployment and configuration, simplifies management complexity, and improves operational efficiency.

Green and energy-saving, reducing TCO: Excellent power-saving ability, on-demand hard drive startup, helping users save nearly 80% of power consumption, while isolating resonance and extending disk life.

HAOSTOR-H100 adopts ultra-high density performance storage servers. For archiving and data analysis of streaming files, large-scale objects, or block level data, a single node supports up to 78 3.5 "SAS/SATA hard drives with a capacity of up to 1PB. The hardware adopts an innovative screw free hard drive carrier design, allowing operators to complete system assembly in a short time, greatly reducing deployment and service time.

Applications

1. High performance computing

The fields of high-performance computing include life sciences, scientific research and education, energy geology, and space exploration. High performance computing applications involve massive data operations and have high requirements for the performance, capacity, and stability of storage systems.

Typical application scenarios: meteorological oceanography, gene sequencing, energy exploration, remote sensing satellites, etc

2. Data protection/safety production

With the internet-based business of many industries, the demand for low-cost storage, backup and archiving of massive data is growing. The HAOSTOR-H100 product provides multiple data protection functions to help multiple data centers build backup archiving, data protection, and disaster recovery capabilities. The system provides performance close to online storage at a cost and power consumption close to offline storage, providing a very suitable new platform for bringing a large amount of offline data to life and use.

Typical application scenarios: financial electronic bill imaging and dual recording (audio/video), medical imaging, vehicle networking, enterprise resources, etc