Saturday, March 21, 2015

Date's Twelve Rules for Distributed Database Systems - Replication independence

Date's Twelve Rules for Distributed Database Systems - Replication independence / Replication transparency


Although the same data item may be replicated (duplicated) at several nodes/sites in a network, with replication transparency, the developer (or other end user) may treat the item as if it were a single item at a single node.
Note: objectives 4 to 6 are helping in ensuring security to certain level.
Example:
Assume that the table following EMP is replicated at all sites. That is, the copy of EMP table is maintained at all the sites.


ENo
EName
DNo
Salary
E1
Sanjeev
3
10000
E5
Meghala
5
7000
E4
Princy
3
15000

If the user writes a query [for example, SELECT * FROM Emp WHERE DNo = 3;], the query will be forwarded to the site/server that is near to the location of the user and executed by that server.

User View:
User will view as a single centralized server and single EMP table as shown in the picture below.
Figure 1 - User view

System View:
According to the system, the table EMP is actually replicated (duplicated) and available at all the sites of the DDBS according to the configuration.
Let us assume that the user generated the query near Mumbai site. The DDBS has to forward the request to the nearest copy of EMP table. That is, the table has to forward the request to Mumbai site.
Figure 2 - System view - request is forwarded to the nearest site

Note: Replication increases availability of data.



Replication independence is desirable for two reasons:
       1. Applications can operate on local copies instead of remote sites.
       2. At least one copy available 






No comments:

Post a Comment