TWELVE RULES for DISTRIBUTED DATABSES (from Date, 19871)

 

I.    Local autonomy. Local data is owned and managed locally, with local accountability and security. No site depends on another for successful functioning.

II.    No reliance on a central site. All sites are equal, and none relies on a master site for processing or communications.

III.    Continuous operation. Installations at one site do not affect operations at another. There should never be a need for a planned shutdown. Adding or deleting installations should not affect existing programs or activities. Likewise, portions of database should be able to be created and destroyed without stopping any component.

IV.    Location independence (transparency). Users do not have to know where data is physically stored. They act as if all data is stored locally.

V.    Fragmentation independence (transparency). Relations between data elements can be fragmented for physical storage, but users are able to act as if data was not fragmented.

VI.    Replication independence. Relations and fragments can be represented at the physical level by multiple, distinct, stored copies or replicas at distinct sites, transparent to the user.

VII.    Distributed query processing. Local computer and input/output activity occurs at multiple sites, with data communications between the sites. Both local and global optimization of query processing are supported. That is, the system finds the cheapest way to answer a query that involves accessing several databases.

VIII.    Distributed transaction management. Single transactions are able to execute code at multiple sites, causing updates at multiple sites.

IX.    Hardware independence. Distributed database systems are able to run on different kinds of hardware with all machines participating as equal partners where appropriate.

X.    Operating system independence. Distributed database systems are able to run under different operating systems.

XI.    Network independence. Distributed database systems are able to work with different communications networks.

XII.    Database independence. Distributed database systems are able to be built with different kinds of databases, provided they have the same interfaces.
 

1Date, Chris. (1987). An Introduction to Database Systems. Vols. I and II. 4th ed. Reading, MA: Addison-Wesley Publishing Co.