Replication
High availability and data protection through mirroring and replication
Redundant dual-controller storage systems or servers in an HA cluster assure operability in the event of hardware failure. Not covered are:
- Human errors
- Software errors / Logic errors
- Water damage
- Overheating (loss of air conditioning), fire
- UPS failure
- Theft
To assure a continuous availability of data and business applications require the implementation of additonal concepts.
Asynchronous replication
The asynchronous (remote) replication updates from the primary site separately kept copies of the data without affecting the local performance over a simple IP connection to the secondary site.
This variant of replication is particularly suitable for disaster recovery sites, where each site can serve as a recovery site for the other. The replication can also include multiple locations that replicate each other. As the replication is asynchronous there is no need to wait for a response from the remote locations. Changes are sent there after a (short) time lag. The size of the delta can be adjusted individually.
Benefits:
- Even at low bandwidth and high latency replication is possible
- easy rollback in case of logical errors
- multiple sites can protect each other.
- Replication also to long distance locations (eg: another continent)
- can speed up the backup
In order to achieve the highest availability, it is advisable to combine this type of redundant data storage with a synchronous mirror and other mechanisms such as snapshots.
Synchronous Mirror
The synchronous (remote) mirror provides fully redundant data storage. With this type of replication,the data in your storage network is written on a second, remote system in real time. This ensures that the data at both locations are always kept in sync. In the event of a system failure or when an entire site fails all data are stored in the other location and can be made available again without loss.
Benefits:
- Increases availability
- No data loss when switching to second data set
- Active load balancing between two sites possible
Synchronous does not mean available
The difference in the available solutions for synchronous data mirroring is clearly in the availability in case of failure. Switching between the synchronized sites must only take a few seconds, to prevent malfunction due to failure and changeover times. Not every system is suitable for this purpose.
Take care, with our help, that by adopting a synchronous mirroring system not only the data safety will be in place but also:
- the availability of the complete infrastructure increases
- the performance is guaranteed
- the ease-of-management is enhanced
- operating costs are reduced
- the environment is future-proof
Our experts will know which is the right environment for your needs and infrastructure.
Ask us:
Tel.: +49 7071/703-400
E-Mail: storage(at)transtec.de
