Page 160 - DCAP402_DCAO204_DATABASE MANAGEMENT SYSTEM_MANAGING DATABASE

Unit 9: Transaction Management

Because recoverability is essential, such a modification must be used for the timestamp protocol Notes
to be practical. Given the added overheads this entails, on top of the (considerable) cost of
maintaining read and write timestamps, timestamp concurrency control is unlikely to beat
lock-based protocols in centralized systems. Indeed, it has mainly been studied in the context of
distributed database systems.

Task If any transaction encounter unexpected situation then transaction
complete of incomplete.

9.7 Implementation of Isolation

Users are responsible for ensuring transaction consistency. That is, the user who submits a
transaction must ensure that when run to completion by itself against a ‘consistent’ database
instance, the transaction will leave the database in a ‘consistent’ state. For example, the user may
(naturally!) have the consistency criterion that fund transfers between bank accounts should not
change the total amount of money in the accounts. To transfer money from one account to
another, a transaction must debit one account, temporarily leaving the database inconsistent in
a global sense, even though the new account balance may satisfy any integrity constraints with
respect to the range of acceptable account balances. The user’s notion of a consistent database is
preserved when the second account is credited with the transferred amount. If a faulty transfer
program always credits the second account with one dollar less than the amount debited from
the first account, the DBMS cannot be expected to detect inconsistencies due to such errors in the
user program’s logic.
The isolation property is ensured by guaranteeing that even though actions of several transactions
might be interleaved, the net effect is identical to executing all transactions one after the other in
some serial order. For example, if two transactions T and T are executed concurrently, the net
1 2
effect is guaranteed to be equivalent to executing (all of) T followed by executing T or executing
1 2
T followed by executing T . (The DBMS provides no guarantees about which of these orders is
2 1
effectively chosen.) If each transaction maps a consistent database instance to another consistent
database instance, executing several transactions one after the other (on a consistent initial
database instance) will also result in a consistent final database instance.
Database consistency is the property that every transaction sees a consistent database instance.
Database consistency follows from transaction atomicity, isolation, and transaction consistency.

9.8 Testing for Serializability

Example: The schedule shown in the following figure is serializable.
T1 T2
R(A)
W(A)
R(A)
W(A)
R(B)
W(B)
R(B)
W(B)
Commit
Commit

LOVELY PROFESSIONAL UNIVERSITY 153