Backup is the foundation of data protection, but any backup means used on the local side cannot avoid the data loss caused by natural disasters or other man-made errors, so offsite redundancy is an important means of protecting data.
In essence, offsite provisioning is based on a copy of the local-side data and sent to a remote site for storage. From the point of view of data replication, off-site backup can be divided into 3 types: host-side standby, network-side standby and storage-side backup. Three kinds of remote mechanisms have advantages and disadvantages host-side redundancy is the installation of replication software on the local host, replicating data to a remote host, which is used to retrieve data transmitted from the network through a special gateway device and to replicate a copy to a remote device; The storage-side backup is the remote replication capability of the local storage device, Copy the data from the storage device to another storage device on the remote. Each of the three remote replication methods has advantages and disadvantages.
The advantage of host-side remote replication is that users can not consider the hardware differences between the endpoints, as long as the software is compatible with the transport protocol. However, the problem is that users must install replication software agents for each host, as well as the compatibility of replication software with different platforms, and the replication software agents that run on the host for a long time will have a certain impact on system performance.
The advantage of network-side replication is that the user does not have to install any software on the host, as long as a network device is added to the network, all the data transmitted in the network will be stored remotely. However, the user must change the architecture of the network, and the data processing capability (throughput) of this network device will also affect the operation of the entire network. In addition, a different device is usually required to receive data at a distance, which means that the user must purchase at least two identical devices.
Storage-side replication is the direct use of the remote replication capabilities of local storage devices to replicate data from the disk directly to another disk device on the remote computer. It has the advantage that the user does not have to install any software on the host, do not need to change the network architecture, as long as the remote replication destination on the storage device path, the operation of the architecture and deployment than the host-side or network-side simpler. But the problem is that most storage-side remote replication mechanisms will require that the two endpoints of replication must be the same or even the same type of storage device, which means that users must purchase two sets of identical storage equipment at a time, instead of purchasing different levels of storage equipment depending on the importance of the primary station and the station. This will bring cost pressure to the user. Storage virtualization Realizing remote backup storage-side remote replication has a simple architecture advantage, but many users are deterred by the need to buy two identical devices. To address this challenge, users can consider using the intermediary of storage virtualization to resolve remote replication issues across different storage devices.
Storage virtualization can be viewed as a storage-bottom infrastructure that consolidates the entire storage environment through virtualized logical tiers and provides services for the various access requirements of front-end servers.
Virtualization products Connect different vendors ' disk arrays, consolidating these different storage devices into a single, unified storage System, which is then allocated to servers that require capacity on the front end. All storage resources are unified under the integration of virtual layer, the connection between front-end server and back-end storage device, from traditional fixed address connection and space image to dynamic connection through virtual layer, which will facilitate the data replication among different storage devices and the integration of multiple storage protocols.
Under the coordination of the virtual layer, the front-end application server will not directly contact the storage device entity, the capacity space provided by the virtual layer is the same for each server on the front end, as long as the user establishes the data synchronization or replication mechanism in the virtual layer, it can use two different storage devices to complete the high availability off-site provisioning.
For the integration of storage environment, users consider the coexistence of different storage protocols, considering the integration of different vendors ' devices. Because of the cost, in addition to using the high price Fibre Channel in the critical application system, the Ethernet network is widely used as the storage transmission channel.
Using storage virtualization products, which support different access protocols, users can use different vendor storage devices (even if they use different transport protocols) at the two endpoints where remote replication is performed, thereby further increasing the flexibility of the storage system.