Data Distribution Strategy:
Many organizations today have distributed computing networks. For these
organizations, a significant problem in physical database design is deciding at
which nodes (or sites) in the network to physically locate the data.
The basic data distribution strategies are:
| Centralized |
All data are located at a single site. It is easy to implement but
it has at least three disadvantages.
- Data are not readily accessible at remote sites.
- Data communication costs may be high.
- The database system fails totally when the central system fails.
|
| Partitioned |
The database is divided into partitions (fragments). Each partition
is assigned to a particular site. Major advantage of this strategy is that data
is moved closer to local users. |
| Replicated |
Full copy of database is assigned to more than one site in the
network. This approach maximizes local access but creates update problems. |
| Hybrid |
In this strategy, the database is partitioned into critical and
non-critical fragments. Non-critical fragments are stored at only one site,
while critical fragments are stored at multiple sites. |
|
