I am confused between clustered and replicated databases.
What is the basic difference between them and when are they used?

Thanks.

On Sat, 06 May 2006 18:35:47 -0700, raghu wrote:

There are several definitions for 'clustered database', depending on the
vendor. However, in general I would say the differences between clustered
and replicated:

clustered: each computer accesses the one copy of the data directly
replicated: each computer has it's own copy of the data

In Oracle's clustered database, all computers actually access the same
disks and therefore have access to all of the data. Other clustered
databases may elect to partition the data across computers and have a way
of sharing that from the 'owning' computer, making it look like each
computer has access to the single data set.

Replication requires copying and storing the data across all interested
computers.

Each has it's own benefits, drawbacks, deterrents and proponents.

raghu wrote:
Two different concepts that bear essentially no relationship
between each other.

In clustering (RAC) you have multiple instances connected
simultaneously to a single database. There is a single copy
of the datafiles being accessed by multiple servers (nodes).

In replication, whether with stand-alone or RAC you are
replicating the data from one database (set of datafiles)
to another.

Daniel Morgan
University of Washington
www.psoug.org