Just today morning I had the chance to participate for a presentation on database replication. It seemed to me like a good topic to blog about today, until I came across the Data Mirroring technique. Even though both these techniques are almost similar to the unobservant mind and are frequently used interchangeably, a subtle yet strong difference makes them apart conceptually.
Data replication and mirroring are both transparent ways of making the database server more tolerant to faults. Such methods are essential to ensure the consistency of business data. However, as the following figure, shows, how these two methods are quite different from their ways of preserving data.
In Mirroring, a single database server maintains a copy of a specific dbspace on a separate disk. Such a remote mirrored data volume comprises of two identical copies of the data connected by Fibre Channel. Both sides of a mirror process read and write I/Os (Inputs/Outputs) to ensure that each copy is a real-time duplicate of the other. The mirror copies are kept in separate data centers, connected over a local- or metro-area network (LAN or MAN).This mechanism protects the data in those mirrored dbspaces against disk failure because the database server automatically updates data on both the disks and automatically uses the other disk if one of the dbspaces fails.
In Data replication,it duplicates all the data that a database server manages on an entirely separate database server (not just specified dbspaces). As this method involves two separate database servers, it protects the data not just against disk failures, but also against all types of database server failures, including a computer failure or the catastrophic failure of an entire site. The source and target servers used for the process of replication are usually separated by a significant distance to safeguard data from disasters that effect a specific geographic location, such as a region-wide power outage.
Two principle modes of replication techniques:
- Asynchronous replication - the primary and secondary data volumes are no more than a few milliseconds out of sync, so the replication is nearly real-time.
- Synchronous replication - the primary and secondary copies are always identical, so it provides a true real-time duplication
Hence, even though like replication, remote mirroring also uses redundancy to guarantee data availability, the techniques should not be confused with each other.
No comments:
Post a Comment