X-EBS is an enterprise grade distributed block storage solution provided by XSKY. It is based on mainstream open-source scale-out storage system Ceph, seamlessly adapting various open architecture hardware and Linux distributions. X-EBS integrates enhancements of performance, reliability, scalability, and manageability on original open-source version, and implements 24*7 ready self-service portal. X-EBS both helps telecom carriers and CSPs build tens of PB scale storage pool, and enterprise get a cost efficient medium to high class scale-out block storage, converging or replacing traditional storage appliances. It supports virtualization, private cloud, database and other OLTP/OLAP applications.
Traditional distributed storages only resolves issues of capacity and scaling-out limitations in proprietary storage appliances. They cannot fully replace proprietary storage appliances in short of narrow applicable scenarios, administration complexity, and poor TCO results. Enterprise user needs better software-defined-storage solution to meet requirements of standardizing and automizing data-center persistence infrastructures other than only for big volume cold data processing.
X-EBS adapts to kinds of deployment architectures. It supports mainstream storage access interfaces for enterprises, both converged and non-converged deployments of application/virtualization. Meet customers' needs of capacity, performance, and manageability in flexible ways.
Based on Ceph core, with performance enhancements and flash/networking optimization.
Realtime QoS setting on volume, adapting to seasonal application.
Redundant design in IO path and management module, no SPF. Multiple data copy with automatic recovery.
Support iSCSI, SCSI, and Fiber Channel. Applicable to major applications and help enterprise re-use procured appliances.
Fine-grained control on data recovery when storage cluster facing permanent or temporary inconsistency.
Impact of malfunctions on software and hardware can be quarantined to a Protection Domain. Other parts of the storage cluster are guarded.
Self learning Read/Write Cache, flexible tiering. With redundancy policies, reliability and performance are well balanced.
Visualized management and automatic maintenance, covering installation, administration, monitoring, and upgrading.
Support EC, snapshot, deep clone, online resizing volume.
Provide enough storage bandwidth and latency characteristics for most OLTP and OLAP database. Support multi-path IO and cluster filesystem for DB HA.
Support VDI and server virtualization based on KVM, XEN, and ESXi, with optimization for performance and reliability. Adapt to mainstream OpenStack distributions. Deploy both in hyper-convergence and storage pool.
Optionally integrate with HadoopFS. User can utilize Map-Reduce paradigm for data analytic tasks, reducing data migration cost between sites.
Support Docker and other container technologies, help user build data availability feature on container based private cloud.
16PB data in single cluster.
1 Volume of maximum 1PB size.
1024 servers in single cluster.
99.999% reliability (Data service available time).