Concurrency Control INFS2200/INFS7903

These flashcards cover key terms and concepts related to concurrency control in relational database systems.

Concurrency Control

Defines protocols and algorithms to manage problems caused by concurrent transactions such as lost update and incorrect summary.

Two Phase Locking (2PL) Protocol

A locking protocol that dictates the positioning of locking and unlocking operations to guarantee conflict serializability.

Wait-for Graph

A graph used to detect deadlocks where nodes represent transactions and edges indicate waiting requests for locks.

Binary Locks

A type of lock that can either be locked (1) or unlocked (0) and is used to control access to data items.

Read/Write Locks

Locks that allow transactions to either read from or write to data items with different states: read-locked, write-locked, or unlocked.

Deadlock

A situation where two transactions are waiting for each other to release locks, preventing both from proceeding.

Livelock

A state in which a transaction is actively changing states but cannot make further progress because it’s continually waiting.

Timestamp Ordering (TO)

A concurrency control method that uses timestamps to determine the order of transaction execution to ensure serializability.

Conflict Serializable Translations

Transactions that can be rearranged into a serial schedule without affecting the final state.

Thomas Write Rule

A rule stating that if an old transaction attempts to write to a data item and its timestamp isn't older than the read timestamp, the write is ignored.

Conflict Serializability

A property of a schedule where the results are equivalent to some serial schedule of the transactions involved.

Lock Manager

A subsystem within a DBMS that maintains and controls the locking mechanisms of transactions.