FileStore unified storage-Shenzhen Kunlunfeng Trading Co., Ltd.

The new generation unified storage

converged storage system is a software-defined distributed storage product that adopts decentralized Shared-Nothing software system architecture and cloud native technology packaging. It can be deployed on general servers (x86 chips or domestic chips), both internally and externally. The network uses TCP/IP networking, abandoning proprietary hardware lock-in and old storage network technology. The system can smoothly expand from TB-level small-scale storage capacity to PB-level large-scale storage capacity, greatly simplifying the maintenance and management complexity of the storage system, and certain Optimize storage system total cost of ownership (TCO) to a greater extent. GStore integrates four storage forms: block, file, object and big data, and provides a rich data interface. Various storage forms share a unified storage resource pool. You can allocate and expand, manage and maintain the required storage forms according to business needs.

GStore provides iSCSI, RBD, and CSI block storage interfaces, is compatible with VAAI and CSI application ecology, and is suitable for typical business scenarios such as virtualization (OpenStack, VMware), containers, cloud desktops, and OLTP databases. It supports iSCSI multipathing, volume snapshots, and volume cloning. , automatic thin provisioning and other enterprise-level storage features to meet the requirements of business continuity, data protection, and storage resource utilization improvement.
GStore unified storage is compatible with the S3 interface and is suitable for large-scale unstructured data storage scenarios such as audio/video, bill images, medical images, backup and archives, radio and television integrated media, backup and archives, etc. GStore unified storage buckets can aggregate multiple storage pools without the user being aware of it, allowing a single storage bucket to span physical pool boundaries, providing users with nearly infinitely expandable single bucket capacity, improving single bucket concurrent access capabilities, and achieving higher aggregation performance. GStore unified storage supports data storage classification, data life cycle management, multi-site global namespace, asynchronous data replication, intelligent data processing framework, data storage encryption, WORM mode and other features.
GStore supports NFS v3/v4, SMB v2.1/v3.0/v3.1/FTP file interface protocols, and also provides high-performance FUSE POSIX file interface, which can be widely used in audio/video, high-performance computing, video surveillance, etc. Business scene. GStore file storage also provides features such as on-demand horizontal expansion, file snapshots, directory-level quotas, and hierarchical data flow.
GStore is compatible with the HDFS big data storage interface and provides computing and storage separation solutions in big data analysis scenarios. It not only meets the efficient processing of massive data, but also helps users achieve true independent expansion of computing and storage resources as needed.

Features

Self-developed distributed storage engine

Self-developed all-terrain multi-modal distributed storage engine adopts minimalist IO stack and full user mode architecture to achieve extremely high software and hardware resources and network utilization, cross-cloud horizontal scalability, computing power sinking and emerging media oriented Optimized, performance and throughput are increased exponentially.

Solving scale and performance challenges

Through the self-developed metadata service engine and distributed storage engine, traditional object storage products solve the problem of limited number of objects in a single bucket, limited size of a single cluster, insufficient small file processing capabilities, difficulty in cross-site data synchronization and management, and high consumption of system resources. Important shortcomings.

Breaking through capacity and geographic boundaries

Transcending the boundaries of storage media, geographical locations, and platforms, it achieves the integration and unification of unstructured data storage resources across hybrid multi-cloud and edges. It can be scaled horizontally, providing storage capacity from small to nearly unlimited, and creating a unified data infrastructure for users. A globally consistent access experience reduces usage costs and complexity.

cloud native

Adapt to the characteristics and requirements of comprehensive cloud IT infrastructure, fully containerized deployment and microservice management, realize minimalist deployment across hybrid multi-cloud and edge, fault self-healing and intelligent operation and maintenance, perfectly adapt to and support cloud native applications .

Interworking with multiple access protocols

Provides multiple protocol interfaces, including S3 object storage interface, Swift protocol interface, NFS file storage interface and HDFS big data storage interface, and supports the same data to be accessed through multiple protocols, saving storage investment costs and improving the convenience of users using data and real-time.

Hybrid multi-cloud data tiering

Complete data classification capabilities support data flow across storage pools (SSD, SAS, and SATA) within the cluster, and across geographical locations and platforms between the edge and the center (private cloud or public cloud) to achieve intelligent cost optimization.

Multi-level safety and reliability

EC erasure coding and multi-copy mechanism, cross-data center data replication, hard disk and network sub-health detection and self-healing and other data reliability technologies, data transmission encryption and server-side storage encryption, WORM (Write Once Read Many) and other security technologies, In addition, data access logs and operation and maintenance operation logs meet user compliance audit and traceability requirements.

Native big data capability support

Native HDFS protocol interface support allows seamless integration with mainstream computing frameworks and tools in the big data ecosystem. The big data business implements a storage and computing separation architecture, with flexible allocation of computing and storage resources, and significant cost optimization; there is no longer a need for a long data preparation process, the big data engine can directly connect to the ET UOS object storage platform to start big data processing and Analysis makes business more agile and efficient.

Composable architecture supports continuous evolution

Based on the combined architecture and on-demand service model of the hybrid multi-cloud data platform, various service components can be subscribed to realize iterative evolution from a single object storage system in a single data center to a block storage system, a big data storage system, and a container Storage system, NAS file system, data lake platform, data management platform, data service platform and even a complete hybrid multi-cloud data platform spanning data center, multi-cloud and edge.

Customer benefits

Only one copy of data is stored within the enterprise, with intelligent management and cost optimization.

A data platform that can be scaled horizontally and continuously is sufficient to store all enterprise data and provide rich protocol interfaces and data service capabilities.

Easily build, deploy and operate various application systems based on the "hybrid multi-cloud + edge" infrastructure environment

Data “originates on the cloud, grows on the cloud, and is applied on the cloud”

The composable architecture allows enterprises to build their next-generation data infrastructure in a more agile way. They can start from a small scale, closely integrate the construction process with the business, and expand on demand.

Enterprise IT shifts from focusing on infrastructure to focusing on business and data, and shifts from operation and maintenance to providing data services

Build the next generation of intelligent data infrastructure to gain new impetus for business growth.