对象存储_海量数据存储_XEOS对象存储

Object Storage-Born for Mass Data

With the fast development of cloud, big data, mobile internet and social networks, unstructured data has exploded. However, the existing distributed file storage fails to satisfy uses’ requirements due to the performance choke in mass small file, cross-region data share, multi-level folder access, concurrent performance and other scenarios. It becomes a key part to choose a more fitting storage system and architecture in enterprise's digital transformation.

IDC: Data scale will reach 40ZB in 2020 and more than 163ZB in 2025.

Object storage is a simpler and more flexible data storage solution that provides an affordable and comprehensive underlying infrastructure. The advanced data storage and management functions of object storage effectively eliminate the limits on namespaces, folder hierarchy, performance, scalability and concurrency, meanwhile it supports on-demand access across regions and ensures safer and easier management on mass unstructured data.

Why XSKY

IDC《China SDS Market Overview》报告显示，XSKY 已经连续 3 年，名列中国对象存储市场第一，拥有大量的客户案例和完善的企业级对象存储构建经验以及领先的技术特性。

China SDS Object Level Market Overview, 2019

Source: IDC 2019Q4 China SDS Market overview

Product Value

XEOS （XSKY Enterprise Object Storage）builds a cost-effective storage architecture based on industry standard X86 servers with leading storage technology, which is infinitely scalable, continuously online and geographically accessible. It ensures data safety, meanwhile eliminating scale and geography restrictions, reducing the investment of enterprise IT construction and meeting the multi-sourced storage requirements generated by new business forms.

Advantages

Infinite Scale

Eliminate scale and geography restrictions, provide a global namespace and integrate storage systems in different locations on demand to achieve near-read
The expansion of the entire pool enables business non-awareness with no change on the way the user invokes data, avoiding the negative impact of data rebalancing and bucket effect on business
Decoupling of software and hardware, supporting heterogeneous deployment of physical devices to meet hardware device upgrade requirements in the future

Persistent Protection

Reliable offering of three-site to achieve fault domain at datacenter-level and disaster recovery
EC, replication, and fault domain mechanism ensure the overall data safety at node level, rack level and datacenter level
Ensure the overall data safety transfer and access with encryption, multi-version, WORM and others
硬盘和网络亚健康处理，自动识别坏盘、慢盘和网络故障，并自动隔离。

Efficient Access

Excellent distributed algorithms to maximize hardware performance and linearly-increased performance to meet performance growth needs in the future
Packed small files improve the I/O performance of small files and better utilize the space
合理的均衡算法，实现资源使用均衡化，大大提升系统整体的读写性能；
数据按策略进行“热”、“温”、“冷”自动分层并流动，实现高效访问和成本最优。

Smart Ecosystem

面向 S3 用户的图形化 Web 应用程序，帮助客户降低使用 XSKY 对象产品的门槛；
A comprehensive ecosystem, compatible with legacy applications, cloud-native and data mining
Data lifecycle management create value from mass unstructured data
Leverage metadata, custom tags and other information to achieve elastic search and classification, supporting big data, artificial intelligence, machine learning and custom data attributes

图片处理

功能丰富，支持图片缩放、裁剪、格式转换、水印、旋转、质量调节等十几种图片处理功能；
图片处理的全处理流程均在对象存储内部完成，减少业务端的交互逻辑，提高硬件资源利用率；
灵活使用，支持 S3 API 、OSS SDK、AWS SDK、XSKY S3 Console 多种使用方式；
灵活扩展，图片处理服务运行在每个对象存储网关上，灵活高并发负载的同时满足服务高可用。

Typical Scenarios

High Concurrency of Small Files

Flat namespace with no performance choke by "directory tree", effectively improving the access efficiency to large files and easily realizing fast R/W and search of small files of 100-billion level

Backup and Archiving

对蓝光库、带库进行统一管理，实现光磁一体的解决方案，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。

Radio, TV & Media

Feasible storage pool with on-demand scale-out capability to support streaming media which requires high scalability, high throughput and high availability storage for VOD, live telecast, news, etc.

AI, Big Data

Seamless connection to TensorFlow, Spark and Hadoop, supporting to customize storage with rich APIs and SDK

Mass, Efficient and Value-added Object Storage