Why use replication?
Replication techniques can be used to improve or enhance availability. Using replication makes identical data available at different systems or even at different locations.
Upon system failure, identical data may be accessed through a different system. Performance of information systems is to be enhanced by offering simultaneous reading from different systems.
One of the main features of replication is transparency to the end users. To them, the process of accessing replicated information is identical to their regular workflow, no additional tools or knowledge are needed.
Type of replication
We distinguish two main types of replication to be used between two storage units:
-
Asynchronous replication
-
Synchronous replication
It is key to note that replication usually is available only when using same-vendor systems. Qi ict offers solutions that provide a work-around for this disadvantage by applying a virtual layer. This technology is called 'Virtual Storage', used in SANRAD appliances.
This technique periodically 'copies' data to a system. Usually, a connection with low bandwidth is used.
Through this method, use of bandwidth is confined to a minimum and thus is relatively inexpensive. A disadvantage however, is the fact that data on the remote side of the system are never completely up to date at all times. This is largely due to RTO (Recovery Time Objective) and RPO (Recovery Point Objective) time frames.
Using this technique, the remote side is always completely synchronized with the primary data. RPO and RTO times have less influence, which proves to be a huge advantage. A definite disadvantage is the fact that this method needs a large bandwidth connection. Even more importantly so, hardly any delay on the connection may occur.
Considering these specifics, only dark fiber is fit for synchronous replication on a WAN connection. Within a datacenter, synchronous replication may be performed over the LAN.