Distributed Database Systems Quiz 2

Concurrency Control in Distributed Database Systems

• The main purpose of concurrency control algorithms is to manage concurrent access to shared data in a distributed database system.

Reliability Measures

• The property of reliability measures that ensures once a transaction is committed, its changes are permanent and persistent, even after system failures, is durability.

Distributed Transaction Processing

• The primary function of distributed transaction processing is to manage multiple transactions executing concurrently in a distributed database system.

Data Management

• Data management in a distributed database system aims to handle large volumes of data from different sources in a scalable and efficient manner, particularly focusing on data integration.

Data Replication

• The primary role of data replication in a distributed database system is to maintain multiple copies of data to improve availability and performance.

Integrity Enforcement

• Integrity enforcement in a distributed database system primarily aims to ensure data consistency and correctness.

Timestamp-Based Concurrency Control Algorithm

• In a timestamp-based concurrency control algorithm, the serialization order is established by assigning a unique timestamp to each transaction.
• The preferred method for timestamp assignment in a distributed DBMS is to maintain uniqueness by using a combination of a unique site identifier and a local timestamp.
• The two-tuple form of a timestamp in the context of assigning timestamps based on local counters is (site-id, local-timestamp).

Basic Timestamp Ordering (TO) Algorithm

• The basic timestamp ordering algorithm tries to achieve a serialization order that is consistent with the timestamp order.
• According to the TO Rule, operation Oij is executed before Okl if the timestamp of Oij is less than the timestamp of Okl.
• When a scheduler encounters a new operation, it executes it if its timestamp is the smallest among all pending operations.
• Assigning read and write timestamps to each data item in the basic TO algorithm ensures that the system maintains a consistent serialization order.

Drawbacks and Enhancements

• A drawback of the basic TO algorithm is that it is prone to starvation, where a transaction may be indefinitely postponed due to higher-timestamp transactions.
• The conservative TO algorithms attempt to lower the possibility of starvation by allowing operations to proceed if they do not conflict with already committed transactions.
• Instead of causing operations to wait, conservative TO algorithms execute them as soon as possible while maintaining a consistent serialization order.
• According to the conservative TO algorithms, each operation is delayed until its timestamp is smaller than all other timestamps of operations that have been executed or are currently executing.
• Conservative TO algorithms aim to achieve a higher level of concurrency while maintaining a consistent serialization order.