Distributed Database Design

Distributed Database Design Exercise

A bank has its head office in Stoke. It has two branches, one in Longton and the other in Burslem. Each customer account belongs to one branch only. Currently, the company has one database located in the head office. Customers at branches access this database via a communication network for whatever they need.

One of the relations in this centralised database system is the Account relation, where data about the accounts is kept. The attributes of the Account relation are: the account number (Acc no), account type (Acc type), the customer’s identification (assume Name), the branch where the account is held, the current balance for the account (balance) and an indicator which tells head office if the publicity (information on updates to regulations, interest rates etc) for a particular account type for a customer has been sent out. An instance of the account relation is given below:

Acc No Acc Type Name Branch Balance Publicity
191688 002 Jones Burslem 2000 yes
779865 005 Smith Longton 600 no
158756 002 Green Longton 8 yes
125467 007 White Longton 10000 no
124678 005 Black Burslem 150 yes


The bank has decided to move to a distributed database system where each of the sites has its own database. Propose a fragmentation design of the Account relation that reflects the distribution of the company’s sites and their functionality. Justify your proposal.