Chapter 8 · Watch, then practise
Transaction and Concurrency Control
Separate atomic commitment across participants from isolation between concurrent transactions. Practice a serialization anomaly and explain why coordination protocols need recovery procedures.
3 questions · 3 with related videos. Matches are based on playlist titles; broader background matches are labeled.
What to study
- Serializable execution
- Concurrent transaction anomalies
- Distributed commit
Chapter playlists
Choose a playlist
Notes
Lec-96: Introduction to Serializability | Transactions Concurrency and Control | DBMS
Gate Smashers · 9:17
Introduces serializability and transaction interleavings.
1. Serializability
Must serializable transactions execute physically one at a time?
No. They may overlap, but successfully committed results must correspond to some serial execution. A database can reject a conflicting transaction to preserve this property; an application must be prepared to retry the whole transaction.
Lec-87: Introduction to Transaction Concurrency in HINDI | Database Management System
Gate Smashers · 12:42 · Background lecture
Transaction-concurrency background; the doctors’ write-skew scenario is explained in the answer.
2. Write-skew example
Two doctors are on call. Each transaction sees the other on call and sets its own doctor off call. Why is the resulting state problematic?
Both can leave no doctor on call, although either transaction alone would preserve one. Separate row updates do not remove the shared predicate dependency. Serializable isolation or suitable explicit coordination must prevent both conflicting transactions from committing this result.
DS38: Distributed Transactions: Atomic Commit protocols| two phase commit in distributed database
University Academy · 9:30
Choose a video · 2 lectures
Two-phase commit lectures support the atomic-commit purpose and blocking discussion.
3. Commit versus consensus
What is the purpose of distributed transaction commit, and why can classic two-phase commit block?
Participants must agree whether the transaction commits or aborts. In two-phase commit, a participant that has prepared may have to wait if it cannot discover the coordinator’s decision. Fault-tolerant commit protocols add mechanisms to recover agreement; atomic commit and isolation solve different problems.