Freely collect, store, compute, manage, and utilize data
Cloudification of storage infrastructure, aggregation management, safe and reliable data storage, and long life cycle management
System Overview
Evergreen Data Operating System ( XSKY DATA OS )
The mission of XSKY is to make data come alive (MAKE DATA ALIVE) by enabling government and enterprise customers to freely collect, store, compute, manage, and utilize their data. That is, the cloudification of storage infrastructure, aggregation management, safe and reliable storage of data, and long life cycle management are realized for the utilization of computing power and innovative applications. Software-Defined Storage (SDS) is architected to keep up with the exponential growth of data while also enabling an automated, agile and economical infrastructure. The software-defined storage provided by XSKY is built on a truly distributed architecture and uses modern distributed system technology to meet all your primary storage, secondary storage and cloud-native data needs.
Evergreen Data OS (XSKY DATA OS) is the base of XSKY's software-defined storage product line, providing a unified runtime environment for all products. The evergreen data operating system can convert ordinary commercial x86 or Xinchuang servers into a storage cluster that can be expanded from a few nodes to thousands of nodes. The advanced software architecture simplifies all aspects of storage and has complete enterprise data functions. Fine-grained configuration at the application level and automation through a complete RESTful API.
Core Advantages
A software-defined storage platform that supports all protocols
By building a distributed system design, complete enterprise-level functions and cloud-oriented architecture simplification, Evergreen Data OS is a rare software-defined storage platform that supports full protocols, and by using different types of storage nodes, it can simultaneously operate at high Performance primary storage and secondary storage areas are used. The overall platform provides the following innovations:
Full-service active-active to meet the demanding requirements of carrying key services
- RTO≈0, RPO=0, zero data loss
- Not limited to block storage services, but also supports file and object services
One pool of multi-core, smooth transition to Xinchuang ecology
- Support rolling upgrade of heterogeneous nodes, no data migration required
- Compatible with mainstream Xinchuang servers and operating systems, performance comparable to x86
Full-scenario EC with both performance and capacity
- XSpeed "unbreakable" Cache, providing performance comparable to replicas
- Can support block EC without full flash
- Support for QLC media and online data reduction
Scientific and technological innovation has produced dozens of national invention patents
- Has applied for more than 70 core technology patents
- Nearly 20 core technology patents have been authorized
Logical Architecture
Evergreen Data OS is committed to providing a unified platform for any application scenario
Evergreen Data Operating System is committed to providing a unified platform to provide a complete storage solution for any application scenario and any deployment architecture. Through a three-tier distributed storage architecture, it supports blocks, files, objects, and big data to provide services based on the same storage pool and perform unified management. As shown in the figure below (the part inside the dotted line box):
Access Gateways and Protocols
Responsible for receiving and processing access to various services, providing standard block protocol ( iSCSI / FC ) support for structured data, and RBD protocol support for KVM scenarios. Provide comprehensive protocol coverage and interoperability for unstructured data
Data Service
Provide corresponding advanced data service support according to the three protocol planes of block system, object system and file system
Indexing Service
The object index pool provides massive metadata support for object systems and corresponding data service capabilities, and supports hundreds of billions of scalability. The file metadata cluster provides high-performance metadata services for file systems and their data services
Distributed Storage Pool
Provide highly reliable and scalable persistent storage services. Responsible for space allocation and exception management of storage media such as HDD and SSD, and implement flexible data distribution strategies through data redundancy algorithms across node devices and cluster physical topology information perception
Custom Operating System
The evergreen data operating system comes with a security-hardened XSKY deeply customized operating system to achieve end-to-end highly reliable services. It also supports the mainstream operating systems in the industry, and is deeply integrated for Xinchuang operating system, which can be adapted to x86 servers and Xinchuang servers at the same time
Storage Management
Provide resource configuration, system management and daily operation and maintenance of the storage system through an intuitive GUI and REST API / CLI
Deployment Architecture
Evergreen Data OS
provides flexible deployment methods
The evergreen data operating system provides a flexible deployment method, allowing each node to define its role according to the storage usage plan. It can be used as a storage node, or as a monitoring and management node. In addition, benefiting from the openness and compatibility of the system architecture, nodes can also be deployed The computing nodes of cloud platforms such as OpenStack / ZStack have become hyper-converged deployment forms.
From the figure on the right, we can see that different role type nodes are deployed with different components, and services can be described according to the logical architecture:
*Deployment View Examples
| Type | Service Name | Describe | Deployment Requirements |
|---|---|---|---|
| Data Service | XDC ( XSKY Data Client ) |
Provide block storage protocols used by structured data, including iSCSI, FC, SCSI, RBD | • ISCSI / FC: 1 on each gateway node • Local SCSI / RBD: 1 on each computer node |
| XGW ( XSKY Object Gateway ) |
Provide object storage service gateway, support S3 and NFS protocols | Each gateway node deploys 1 | |
| XPG ( XSKY Parallel Gateway ) |
Provide file storage service gateway, support CIFS, NFS, FTP, S3, HDFS protocols | • NFS / SMB / FTP / HDFS: 1 per gateway node • POSIX: 1 per compute node |
|
| Indexing Service | XMDS ( XSKY MetaData Service ） |
Provide file metadata indexing service | Deploy at least 3 to form an XMDS cluster |
| XINDEX ( XSKY Object Index ) |
A key-value storage service that provides object indexes | Deploy at least 3 to form a XINDEX cluster | |
| Distributed Storage Pool | MON ( Monitor ) |
Provides the operation monitoring and configuration management of the cluster data plane, and manages the data distribution strategy of all storage pools | Deploy at least 3, you can choose 3, 5, 7 |
| OSD ( Object Storage Device ) |
Provide data persistence capability, each storage medium has a corresponding OSD service process | Each storage node is deployed, and the number of deployments is the number of storage media of the node | |
| Storage Management | XMSD ( XSKY Management Service Daemon ) |
Provides cluster deployment, management, and operation and maintenance functions, and distributes them on each node. More than 3 nodes can be set as master management nodes, responsible for management data maintenance and global monitoring | Deploy 3 to form a management service cluster |
| DPGD ( DATA Protection Gateway Daemon ) |
Provides functions related to data protection, including structured data backup, asynchronous replication, and replication relationship management.Only start if data protection is enabled | Deploy when creating a protection gateway |
