Key Features

• Open solution, easy to expand and integrate
• Support many types of storage for Object, Block, File Storage
• Storage system has the ability to expand up to PBs
• Ability to run on popular Intel servers, do not depend on hardware vendors
• Ability of self-correction, self-balancing, fault tolerance
• No Single Point of Failure

Benefits

• Use as alternative solution instead of traditional SAN solution with lower cost and infinite scalability
• Implement HA services such as Load Balancing for Web Server, DataBase Replication
• Serve the request of high processing capacity and optimal I/O, thanks to automatic Cache Tier, Storage Tier, Replication, Ensure Coding, Cold Storage for Backup
• Support virtualization, Cloud Storage
• Reduce CAPEX investment costs and OPEX operating costs
• Pay as grow (investment based on the development of infrastructure)
• Support 1Gbps connection platform, 10Gbps, 40Gbps, InfiniBand

​​Solution architecture

CEPH solutions build on software daemons. Each process has its own role in Ceph's features and special values. This is a factor contributing to reduce costs when comparing Ceph with similar systems.
 

• Reliable Autonomic Distributed Object Store (RADOS)
• The basic element creats the Ceph storage cluster. Ceph data is stored on object, RADOS obj is responsible for storing, regardless of data type.
• RADOS data layer will always be accurate and secure. In terms of consistency, it will be replicated, detect errors and recover on every node in the cluster.
• When storage app goes to Ceph cluster, data will be stored in Ceph Object Storage Device (OSD) as object. This is the only component Ceph cluster uses to store data and retrieve it. Typically, the total physical disk in Ceph cluster will equal the number of OSD daemons running to store data to each disk.
• Ceph monitors (MONs)

The component focuses on the whole cluster status, OSD, MON, PG, CRUSH map status monitoring. The cluster nodes will monitor and share information about the changes. The monitoring process will not store data (this is the work of the OSD).
Librados lib supports RADOS access via PHP, Ruby, Java, Py, C/++. Provide a friendly interface to Ceph storage cluster, RADOS and RBD, RGW, POSIX interface services in CephFS.

• Ceph Block Device hay RADOS block device (RBD)
• The component provides block storage, can be mapped, formmatted, mounted as any normal disk. Ceph block device supports provisioning and snapshots.
• Ceph Gateway Gateway or RADOS gateway (RGW)
• Component provides RESTful API interface, compatible with Amazone S3 (Simple Storage Service) and OpenStack Object Storage API (Swift). RGW also supports OpenStack Keystone authentication services.
• Cephet Metadata Server (MDS)
• The component focuses on file hierarchy and stores metadata exclusively for CephFS. Ceph block device and RADOS gateway do not require metadata because they do not need MDS Ceph daemon. MDS does not directly support customers, thus eliminating the single error of the system.
• Ceph File System (CephFS)
• Components provides POSIX-compliant, distributes filesystem for all types. CephFS is based on Ceph MDS to show file hierarchy, metadata.

Customers

• Hanoi television broadcasting
• Direct for road of Vietnam - Ministry of Transport
• Viet Light Technology Joint Stock Company
• Hanelcom Data Center
• ViVAS VNPT