Software Defined Storage AS13000G5

Inspur AS13000G5 is a fully symmetrical distributed enterprise-level storage platform for massive data processing applications. The industry-leading "all-in-one" minimalist architecture can provide four kinds of storage service for file, block, object, and big data on one platform at the same time. AS13000G5 supports the maximum horizontal expansion of 5120 nodes, meanwhile a single namespace can support EB-level capacity, and the performance and capacity of the system vary with the number of nodes linearly.AS13000G5 fully meets the storage and disaster recovery requirements of massive data in cloud computing, virtualization, database, file sharing, bill imaging, HPC, video surveillance, archive backup, etc. and has been widely used in government, finance, enterprise, telecom, radio and television, education, medical, transportation, energy and other industries.

Key Features

Minimalist architecture

Inspur distributed storage adopts a fully symmetric and decentralized distributed storage architecture. A single cluster provides customers with four storage services of file, block, object and big data, greatly improving storage utilization and ease of management.

(1) File storage service: AS13000G5 supports standard communication protocols such as CIFS/NFS/FTP/HTTP, supports data sharing and access in different operating systems. provides tiering, enhanced ACL, quota, recycle bin and other functions to meet customers storage needs in different scenarios.

(2) Block storage service: Provides standard block storage based on ISCSI protocol, supports docking with virtualized cloud platforms, and provides functions such as snapshots, cloning, and thin provisioning to provide flexible storage space and reliable data for virtualization and cloud computing

(3) Object storage service: It is compatible with S3/Swift protocol and provides multiple SDKs to meet the data backup and data retrieval requirements of big data services in different scenarios. It supports multi-tenancy, multi-source zero copy, WORM and other features.

(4) Big data storage service: supports HDFS API, effectively solve the problem of space waste of the three-copy data storage mode of Hadoop platform and achieve cross-platform data shared access to meet the needs of different applications.

(5) Distributed database protocol: Support distributed database protocol interface to access storage, use database protocol interface to quickly locate specified files or objects, and manage different attributes of files or objects, which can effectively reduce cluster pressure.

(6) Unified resource pool: Supports NFS/SMB/HTTP/FTP/POSIX/S3/SWIFT and other protocols, providing customers with four storage services of files, blocks, objects and big data, and supports IPV6/IPV4 dual stack networking.

(7) The same storage pool supports file/block/object/ big data services for simplify management. 

Extreme performance

Inspur distributed storage adopts the leading Iturbo engine innovation technology, through small file aggregation, object aggregation, global cache, multi-channel, kernel client and other technologies to provide customers with innovative performance acceleration solutions to comprehensively improve storage system performance.

(1) The performance grows linearly with the nodes, and through extreme horizontal expansion, it provides users with PB-level performance.

(2) Provide cache optimization, small file aggregation, object aggregation, automatic tiering and other performance optimization technologies to greatly improve storage performance.

(3) Support configuring NVDIMM for log acceleration, greatly improving the performance of small files.

(4) Support global read-write cache, improve cache hit rate, and greatly improve read performance.

(5) Support CIFS multi-channel, which can greatly improve business performance and reliability, and use other channels to continue data transmission after a single connection channel fails without affecting the business.

(6) Configure the file-based automatic tiering to achieve intelligent storage and dynamic migration of hot and cold data between different tiering levels.

Extremely stability

Inspur distributed storage provides users with powerful data protection capabilities and system availability through industry-leading redundancy design, with data availability up to 99.9999%.

(1) Support the deployment of 2-8 copies, the maximum erasure is 16+4, and provide advanced erasure (K+M:1) for small scale clusters.

(2) Provide 4TB/h hard disk reconstruction speed, greatly reducing cluster downtime.

(3) Provide AI-based disk failure prediction technology to perform planned HDD reconstruction to reduce the risk of data loss and performance degradation.

(4) Provide triple data protection at the hard disk level, node level, and cabinet level to reduce the risk of data loss.

(5) Provide HDD anti-misplugging service, customer can insert it back within 5 minutes and continue to use it without data reconstruction.

Extreme capacity

Inspur distributed storage provides users with massive storage capacity, and through a series of advanced functions, greatly improves capacity utilization and provides customers with a lower TCO through a leading system architecture.

(1) Support the expansion of 3-5120 nodes and single cluster provides users with EB-level capacity space.

(2) Through extreme node-level and hard disk-level balancing, the capacity error rate between disks does not exceed 5%, the error between nodes does not exceed 1%, and the available capacity utilization rate is as high as 95%.

(3) Storage can support NAS, object and HDFS three storage services to access the same file at the same time with no need to install any client software or plug-in.

(4) Through the extreme architecture design, a single directory supports 10 million files, and the file system supports tens billions of files.

(5) Through extreme architecture design, a single bucket supports hundred billions of objects in the object scenario.

(6) Support the expansion of capacity and performance without service interruption by adding storage nodes to the storage system without downtime. The time to expand a node is no more than 1 minute, and the performance loss is less than 10%.

(7) Support the capacity reduction without removing the storage node by removing storage nodes from the storage system without downtime. The time for scaling down a node is no more than 1 minute, and the performance loss is less than 5%.

Extreme cloudification

Inspur distributed storage provides a standard interface to connect various virtualization platform,containers and cloud platforms to help users achieve efficient data circulation in a hybrid cloud environment.

(1) Support CSI container interface to meet storage requirements under container cloud.

(2) Support Manila, Cinder, Swift, S3 and other protocols to connect various private/public clouds.

(3) Support restful API to connect with various of cloud management platforms.

Extreme manageability

Inspur distributed storage management interface adopts humanized GUI designed to provide users with a simple and easy-to-use operation interface. At the same time, through the multi-device management software InView which support using one unified Web interface to manage multiple sets of clusters, and provide AIOPS function, giving customer an intelligent and efficient operation and maintenance experience.

(1) Combine InView and GUI interface: operation and maintenance visualization, management of multiple storage clusters on the same Web interface without third-party software or plug-ins. 

(2) Support docking with InView, provide AIOPS function including advanced functions such as disk failure prediction, performance trend prediction and capacity trend prediction, the overall prediction accuracy rate reaches 99%.

(3) Support sending alarm information via SMS, email, etc.

(4) Support QoS based on connection number, bandwidth, etc.

(5) Support quota management based on directory, user, etc.

Extreme security

The release of Inspur distributed storage strictly complies with the enterprise-level high-standard security baseline, and has passed thousands of security tests, covering data security, system security, communication security, and application security, providing users with a secure data storage platform.

(1) Support WORM function, provide users with different granularity file write protection.

(2) Support enhanced ACL, providing five different permissions such as read, write, execute, delete, and visual.

(3) Meet the security baseline requirements, provide features such as user authentication, multiple data encryption protection, and web application hardening to ensure the safe use of storage.

Technical Specifications




Storage Service Type



Up 3 to 5120 nodes



Data Redundancy Protection

Multi-copy (2-8), EC K+M:1 (maximum 32+2)

Set data copy and erasure code in the same system according to different data protection levels.


Max file system size: 1024PB

Max files per directory: million files

Max file size:128TB


Default block size4MB

Configured size1MB,2MB,4MB,8MB,16MB


Max Object Size: 48.8TB

Max Objects Per Bucket: No limit

Max Buckets Per User: No limit

Max Objects Per Cluster: No limit

Max Users: No limit

Hardware specification

Model Type




Network Type

1/10/25/40/100Gb Ethernet, 16/32Gb FC, 100Gb InfiniBand


Key Features

Online global deduplication, compression, auto tiering, remote copy, intelligent disk prediction, auto thin-provisioning, QoS, active-active, clone, snapshot, enhanced ACL, multi-tenant, WORM, multi-source zero copy, soft copy, recycle bin, kernel client, etc.


Compatible Platforms and Components

VMware ESXi NAS/iSCSI, VAAI OpenStack and CSI