According to IDC's "Rethink Data" report, only 32% of the data generated by enterprises in operation are utilized. Unstructured data contain a great deal of information, and unstructured data account for 80% of the data generated by enterprises, which is increasing rapidly at a compound annual rate of 60%. The management of massive unstructured data and tapping into their value has become a problem that plagues enterprises during their data management.
Unstructured data are organized in the form of files and objects. The method with files adopts a hierarchical architecture, which is more in line with human logic. Files can be frequently edited and modified for file sharing, video editing, high-performance analysis, etc. The method with objects adopts a flat structure, which is more machine-friendly. Objects are seldom modified and suitable for cloud native application, cloud resource pool, content release, etc. However, the scenarios where files and objects are used are not fixed. Most applications based on file storage can apply to object storage mode after being appropriately transformed. In addition, under scenarios of backup, archiving and big data analysis, either file storage mode or object storage mode can be adopted. Therefore, enterprises also possess cross-protocol capability for data flow to manage unstructured data uniformly and arrange resources reasonably.
XUDS（XSKY Unstructured Data System）是 XSKY 星辰天合打造的一站式非结构化数据管理平台，采用软件定义的架构，支持 POSIX、NFS、SMB、FTP、S3、HDFS、CSI 等协议，让存储摆脱硬件和平台的限制，赋予了存储灵活敏捷的特性，形成对庞大非结构化数据的承载能力，这种承载能力是实现数据资产化坚实的基础。
Easy to use with global namespace. Support file sharing, network disk and other office scenarios
Scale-out, rolling upgrade, and permanent preservation of data.
Compatible with HDFS and efficient metadata processing mechanism for flexible response to AI/ML data analysis requirements.
Support CSI interface and multiple PODs to share data
内置 X3DS（XSKY 星辰天合立体数据管理系统）和 ODPF（开放数据保护框架），同时无缝对接多种备份软件，提供整体的备份、归档解决方案。
Rich interface protocols with capability of extension to thousands of nodes
Excellent data reading & writing and load balancing functions meet the needs of mass data and high concurrent access.