Question

Hoffer Chapter 13

Name the type of transparency (location, replication, failure, concurrency) that is indicated by each statement.

a. End users in New York and Tulsa are updating the Engineering Parts database in San Mateo at the same time. Neither user is aware that the other is accessing the data, and the system protects the data from lost updates due to interference.

b. An end user in Tulsa deletes an item from the Standard Price List at the site. Unknown to the user, the distrib uted DBMS also deletes that item from the Standard Price List in San Mateo and New York.

c. A user in San Mateo initiates a transaction to delete a part from San Mateo parts and simultaneously to add that part to New York parts. The transaction is com pleted in San Mateo but, due to transmission failure, is not completed in New York. The distributed DBMS automatically reverses the transaction at San Mateo and notifies the user to retry the transaction. What if the distributed DBMS remembers the failed transac tion component and repeats it immediately when New York becomes available? What risks would this type of approach create?

d. An end user in New York requests the balance on hand for part number 33445. The user does not know where the record for this part is located. The distributed DBMS consults the directory and routes the request to San Mateo.

Answer 1

a.
Concurrency Transparency: Users and Applications should be able to access shared data or objects without interference between each other.

b.
Replication Transparency: This ensures that the replication of the user database is hidden from the user. Whenever a user updates a data item, the update is reflected in all the copies of the table.

c.
Transaction Transparency must display atomicity: that is all the components of a transaction must be completed and permanently stored in the database to ensure the database's consistency and integrity. If a transaction cannot be executed in its entirety, the transaction must be aborted and completed components rolled back.
The main problem with such system is inconsistency of data. If the server that remembers the pending update may fail, causing inconsistency in the database, of partially updated database. It even revokes the property of atomicity of a transaction.

d.
Location Transparency: The user is not aware of the actual location of the database, and can query on any tables of fragmented tables as if it is present locally in his system.