LST E5000 Series Distributed Storage

LST E5000 Series Distributed Storage

The LST E5000 series distributed storage integrated products launched by LUISUAN Technology provide users with intelligent, efficient, reliable and agile high-performance file storage and object storage services with its excellent self-developed capabilities. The LST E5000 series supports multiple protocol interfaces such as POSIX, CIFS, NFS, HDFS, S3, CSI, etc., meeting the parallel access of HPC and efficient access of big data.

Features

Model

Specifications

Value

Features

Adopting a fully symmetrical distributed architecture, no need to deploy metadata nodes independently
Automatically distribute data evenly, and automatically load balance between nodes
Support online expansion, and can modify data redundancy policies online
A maximum of 4 nodes or disks can fail at the same time without data loss and business interruption
Single-node bandwidth can reach 13GB/s, and IOPS can reach 1.5 million
Data lifecycle management, global unified namespace
Tens of billions of files can be retrieved in seconds, and a single volume can receive millions of IOPS
Diskless (storage and computing separation) architecture, on-demand expansion, linear growth of performance and capacity
Support multiple storage services such as blocks, files, objects, and big data

Model

56190486

Perfect storage and computing decoupling:

Decouple the computing CPU and external memory to form a diskless computing pool and remote storage pool, completely breaking the boundaries of various storage and computing hardware resources.

Optimize hardware resource combination:

Offload storage IO processing, which general-purpose CPUs are not good at, to dedicated processors (ASIC/FPGA), achieving the optimal combination of hardware resources and improving computing power energy efficiency.

Specifications

Fully self-developed high-performance IO software stack	NVMe over RDMA new generation technology	Dedicated hardware chips realize data processing
The LST E5000 series distributed storage uses a high-performance IO software stack fully developed by Green Computing Technology. By reducing IO paths, improving concurrent throughput, reducing thread switching and lock-free mechanisms, the performance of new storage media is maximized.	By using the RDMA (Remote Direct Memory Access) protocol, an ultra-high-performance storage network is provided, and its latency is comparable to direct-connected storage.	The LST E5000 series all-flash storage pool uses dedicated hardware chips (FPGA/ASIC) to realize data processing, with the advantages of high performance, low latency, low power consumption and high throughput.

Value

High performance and low latency

Financial-grade high availability

Elastic expansion and cloud-native adaptation

Intelligent operation and maintenance and cost optimization

Industry scenario adaptability

· Adopt all-flash architecture and RDMA network technology to meet the needs of high-concurrency scenarios of enterprises;

· Optimize I/O performance through SSD cache technology and significantly improve data transmission efficiency.

· Supports multi-node redundancy and clustered deployment, and achieves business continuity through the "two-site three-center" disaster recovery mechanism, with data persistence reaching 99.99999% (7 nines), ensuring 24/7 uninterrupted business operation;

· Natively compatible with virtualization and container platforms such as Kubernetes and VMware, providing persistent storage support for databases and stateful applications in cloud environments, and realizing dynamic resource allocation.

· Supports horizontal expansion from 3 to 4096 nodes, with storage capacity and performance increasing linearly with the increase of nodes, meeting the explosive growth needs of enterprise agile businesses;

· Provides a visual monitoring interface, full-dimensional management, and combines automated data balancing and recovery mechanisms to reduce the complexity of operation and maintenance;

· Supports hybrid storage architecture (all-flash + HDD), balances performance and cost through capacity tiering technology, and adapts to cold data scenarios such as archiving and backup.

· Supports complex business loads such as virtualization, AI training, and big data analysis;

· Compatible with domestic operating systems (such as Kylin and Tongxin) and multi-platform protocols (iSCSI, NVMe-oF), meeting the flexibility and security requirements of enterprise-level data management.