Ship to USA and Canada ONLY

Huawei OceanStor CSS

The Cloud Storage System (CSS) is a powerful distributed systems platform (Distributed file systems and Distributed database) featuring cutting-edge distributed system architecture, large capacity, flexible expansion, and intelligent resource allocation and management. It meets the requirements of file services, websites, file backup, streaming media, video surveillance, behavior analysis, log monitoring, historical data retrieval, and data mining. As an industry-leading cloud storage system, it is widely deployed in fields such as government, enterprise, education, telecom, broadcast & television, and Internet.

Product Features

Database Features

  • Compatible with SQL-92 syntax standards.
  • Supports common statistical analysis querying including Group and Sort statements.
  • Supports multi-table joint querying including Join statements.
  • Supports JDBC and ODBC interfaces.
  • Optimizes storage space and resource utilization with the exclusive data compression technology, which is transparent to applications and can achieve a compression ratio up to 7:1.

File System Features

  • Support POSIX, NFS, CIFS, FTP protocols

High Reliability & Availability

  • Employs inter-node RAID technology — Erasure code, which allows the failure of up to three nodes (66 data disks) at the same time in each domain with fewer than 22 nodes and is completely transparent to service hosts.
  • Uses a full redundancy design to ensure 99.999% availability.
  • Disk Pre-copy Technology, Faulty disks are detected in advance and data on the faulty disks is migrated to normal disks to avoid the raid reconstruction and extend disk life
  • Bad Sector Repair, Repair bad sectors with best effort, reducing the disk failure rate

High Performance and Scalability

  • Employs the all-in-one architecture for centrally managing hardware, networks, and database services.
  • Reduces customers' management costs with the wizard-based Integrated Storage Manager (ISM) software.
  • Improves troubleshooting efficiency with fault monitoring and alarming (by email, SNMP, or Syslog).
  • Supports automatic performance statistics and analysis.
  • Offers a namespace-based deduplication technology for maximum storage capacity utilization and reduced overall costs and management complexity.

Easy Management

  • Uses SSDs for enhanced metadata access performance.
  • Eliminates the bottleneck of standard storage nodes and metadata nodes while enabling cross-node concurrent access and to provide 32GB system distributed cache of each node
  • As the number of nodes in the CSS increases, performance improves linearly, throughput per domain reaches a couple of GB/s, the maximum system throughput reaches hundreds of GB/s, and storage capacity is over 100 PB.
  • Supports capacity- and performance-based load balancing. The data migration that balances load based on the usage of physical resources such as the CPU, cache, network I/O bandwidth helps ensure system performance and keep capacity balance and applies to the computing cluster that has an extremely high performance requirement.
  • Supports online components' hot swapping and nodes' increase and decrease, which is completely transparent to services and implements the real online capacity expansion.

 

Product Specification

Primary Feature

Sub Feature

Description

Distributed File System

System Specifications

System capacity

Supports a maximum of 128 domains of total 185 PB.

Capacity of a single domain

A single domain supports a maximum of 22 storage nodes of total 1.45 PB.

Bandwidth of a single domain

The maximum read and write bandwidths of a single domain are 5.6 Gbit/s and 4 Gbit/s respectively.

Number of clients

A single domain supports a maximum of 64 client connections, whereas the systems support a maximum of 512 client connections.

Number of files

A single domain supports a maximum of 1.2 billion files.

File System Functions and Features

System architecture

Employs the distributed and asymmetrical object storage architecture.

File/Directory management

Supports various basic operations against files and directories, including opening, closing, reading, and writing files as well as creating, deleting, and modifying directories.

Supported protocol

Standard POSIX

File size

The maximum size of a single file is 2 TB.

Data migration

Supports cross-node data migration within the same domain.

Data deduplication

Supports namespace-based deduplication and scheduling.

Load balancing

Supports performance- and capacity-specific load balancing among storage nodes.

Thin provisioning

Supports thin provisioning for file systems to allocate space according to the actually required capacity.

Client Compatibility

Client (POSIX)

Supports SUSE Linux 10 sp1, SUSE Linux 10 sp2, SUSE Linux 11 sp1, and RHEL 5.

NFS client

Supports SUSE Linux 10 sp2 and SUSE Linux 11 sp1.

CIFS client

Supports Windows 2003 and 7.

Distributed database

System Specifications

System capacity

Supports a maximum of eight domains of total 2.816 PB.

Capacity of a single domain

A single domain supports a maximum of 16 storage nodes of total 352 TB.

Minimum configuration of a single domain

Each domain contains a minimum of one management node, three storage nodes, and two switch units. Data services and metadata services are all deployed on the storage nodes.

Performance of a single domain

In the minimum configuration and when the number of table indexes is not more than three, the insertion performance of a single node is not less than 20 thousand insertions per second without copies. If there are copies, the performance varies in accordance with the following formula where N stands for the number of copies: 20000/(N+1).

Number of client connections

A single domain supports a maximum of 64 client connections, whereas Support a maximum of 512 databases.

Number of storable events

A single domain and it can store a maximum of 100 thousand and 800 thousand events respectively.

Database Functions and Features

SQL syntax

It is compatible with SQL-92 syntax standards. Also, it supports the ability to create or destroy databases, tables, and indexes; as well as to query and insert data.

External interface

Supports JDBC and ODBC interfaces and reserves data adding, deletion, modification, and querying interfaces in the SDK for secondary development.

Data compression

Supports data compression at the datatable level with a maximum compression ratio of 7:1.

Common feature

Reliability

Disk-level data protection

Supports disk-level RAID 0 and RAID 5 for data redundancy.

Node-level data protection

Supports cross-standard-node data duplication or multiplication, which can recover the data of a failed node using other nodes.

Metadata node reliability

Supports the intra-domain A/P mode for metadata nodes.

Scalability

Hard disk

Supports online adding and deletion of hard disks.

Standard storage node

Supports online adding and deletion of standard storage nodes.

Domain space

Supports online expansion and contraction of domain space.

Maintainability
Reliability
Scalability

Automatic deployment

Supports automatic deployment and configuration for software platforms.

Online upgrade

Supports online upgrade without interrupting ongoing read/write services nor affecting system settings, logs, alarms, and users' data.

Status monitoring

It can monitor the operating status of the entire system and that of each node.

Performance monitoring

It can monitor system performance including read/write performance, I/O performance, and network bandwidths.

Resource monitoring

It can dynamically monitor system resources including disk space, CPUs, memories, network traffic, and file systems.

Device discovery

Supports device discovery and automatic discovery for newly added devices against user-defined IP addresses, network segments, and subnets.

Remote maintenance
Alerting

Supports remote maintenance using SSH or Web and remote log collection.

Supports GUI-based alerting and various remote alerting modes including SMS and email alerting.

Permission management
Disk-level data protection
Node-level data protection

Supports three levels of management users, namely super administrator, administrator, and read-only user, and supports permission- and domain-specific user management.

Supports disk-level RAID 0 and RAID 5 for data redundancy.

Supports cross-standard-node data duplication or multiplication, which can recover the data of a failed node using other nodes.

Metadata node reliability

Supports the intra-domain A/P mode for metadata nodes.

Hard disk

Supports online adding and deletion of hard disks.

Standard storage node

Supports online adding and deletion of standard storage nodes.

For more details, please refer to Huawei OceanStor CSS Distributed File System Datasheet(PDF).