Tag Archives: Durability

🔄 Transactions & ACID Properties

Leave a reply
Image
Image
Image
Image

📘 1. Introduction to Transactions

A transaction in database systems is a sequence of one or more operations performed as a single logical unit of work. These operations may include:

  • Reading data
  • Writing data
  • Updating records
  • Deleting records

The key idea is simple but powerful:

👉 Either all operations succeed, or none of them do.

🔹 Real-Life Example

Consider a bank transfer:

  1. Deduct ₹1000 from Account A
  2. Add ₹1000 to Account B

If step 1 succeeds but step 2 fails, the system becomes inconsistent. Transactions prevent this by ensuring all-or-nothing execution.

🔹 Formal Definition

A transaction is:

  • A logical unit of work
  • Executed completely or not at all
  • Ensures database consistency

🧠 2. Why Transactions Are Important

Transactions are critical for:

  • Data integrity
  • Reliability
  • Consistency across operations
  • Handling system failures
  • Concurrent access management

🔹 Without Transactions

  • Partial updates
  • Data corruption
  • Lost data
  • Inconsistent state

🏗️ 3. Transaction States

Image
Image
Image
Image

A transaction passes through several states:

🔹 1. Active

  • Transaction is executing

🔹 2. Partially Committed

  • All operations executed, waiting to commit

🔹 3. Committed

  • Changes permanently saved

🔹 4. Failed

  • Error occurs

🔹 5. Aborted

  • Rolled back to previous state

🔁 4. Transaction Operations

🔹 BEGIN TRANSACTION

Starts a transaction.

BEGIN;

🔹 COMMIT

Saves changes permanently.

COMMIT;

🔹 ROLLBACK

Reverts changes.

ROLLBACK;

🔹 SAVEPOINT

Creates checkpoints.

SAVEPOINT sp1;
ROLLBACK TO sp1;

⚖️ 5. ACID Properties

Image
Image
Image
Image

ACID properties ensure reliable transactions:

🔹 A – Atomicity

👉 All or nothing

  • If one operation fails → entire transaction fails
  • Ensures no partial updates

Example:

  • Money deducted but not added → rollback

🔹 C – Consistency

👉 Database remains valid

  • Enforces rules, constraints
  • Moves from one valid state to another

🔹 I – Isolation

👉 Transactions do not interfere

  • Concurrent transactions behave independently

🔹 D – Durability

👉 Changes are permanent

  • Even after crash, data persists

🔍 6. Atomicity in Detail

Atomicity ensures:

  • No partial execution
  • Rollback on failure

🔹 Implementation Techniques

  • Undo logs
  • Write-ahead logging (WAL)

🔹 Example

BEGIN;

UPDATE Accounts SET balance = balance - 1000 WHERE id = 1;
UPDATE Accounts SET balance = balance + 1000 WHERE id = 2;

COMMIT;

If second update fails → rollback entire transaction.

🧩 7. Consistency in Detail

Consistency ensures:

  • Constraints are maintained
  • Rules are enforced

🔹 Types of Constraints

  • Primary key
  • Foreign key
  • Check constraints

🔹 Example

  • Balance cannot be negative
  • Foreign key must exist

🔄 8. Isolation in Detail

Image
Image
Image
Image

Isolation prevents interference between transactions.

🔹 Problems Without Isolation

1. Dirty Read

Reading uncommitted data

2. Non-repeatable Read

Data changes between reads

3. Phantom Read

New rows appear unexpectedly

🔹 Isolation Levels

LevelDescription
Read UncommittedLowest isolation
Read CommittedPrevents dirty reads
Repeatable ReadPrevents non-repeatable reads
SerializableHighest isolation

🔐 9. Durability in Detail

Durability ensures:

  • Data survives crashes
  • Stored permanently

🔹 Implementation

  • Transaction logs
  • Disk storage
  • Backup systems

🔹 Example

After COMMIT:

  • Power failure occurs
  • Data still exists

🧠 10. Concurrency Control

Image
Image
Image
Image

Concurrency control manages multiple transactions.

🔹 Techniques

1. Locking

  • Shared lock (read)
  • Exclusive lock (write)

2. Two-Phase Locking (2PL)

  • Growing phase
  • Shrinking phase

3. Timestamp Ordering

  • Based on timestamps

4. MVCC (Multi-Version Concurrency Control)

  • Multiple versions of data

🔁 11. Deadlocks

🔹 What is Deadlock?

Two transactions wait for each other indefinitely.

🔹 Example

  • T1 locks A, needs B
  • T2 locks B, needs A

🔹 Handling Deadlocks

  • Detection
  • Prevention
  • Timeout

🧪 12. Logging and Recovery

Image
Image
Image
Image

🔹 Types of Logs

  • Undo log
  • Redo log

🔹 Recovery Techniques

  • Checkpointing
  • Log-based recovery

📊 13. Distributed Transactions

🔹 Challenges

  • Network failures
  • Data consistency across nodes

🔹 Two-Phase Commit (2PC)

  1. Prepare phase
  2. Commit phase

🔹 Three-Phase Commit (3PC)

Improved version of 2PC

🌐 14. Transactions in Modern Databases

🔹 SQL Databases

  • Strong ACID compliance

🔹 NoSQL Databases

  • Often use BASE model
    • Basically Available
    • Soft state
    • Eventually consistent

⚖️ 15. ACID vs BASE

FeatureACIDBASE
ConsistencyStrongEventual
AvailabilityModerateHigh
Use CaseBankingSocial media

📈 16. Performance Considerations

  • High isolation → slower performance
  • Low isolation → faster but risky

🔹 Trade-offs

  • Consistency vs performance
  • Isolation vs concurrency

🧩 17. Real-World Applications

🔹 Banking Systems

  • Money transfer
  • Account updates

🔹 E-commerce

  • Order processing
  • Payment transactions

🔹 Airline Booking

  • Seat reservation

🧠 18. Advanced Topics

  • Nested transactions
  • Long-running transactions
  • Savepoints
  • Distributed consensus

🏗️ 19. Best Practices

  • Keep transactions short
  • Avoid unnecessary locks
  • Use proper isolation level
  • Monitor performance

⚠️ 20. Common Issues

  • Deadlocks
  • Blocking
  • Performance bottlenecks
  • Data inconsistency

🔮 21. Future Trends

  • Cloud-native transactions
  • Distributed ACID systems
  • New consistency models

🏁 Conclusion

Transactions and ACID properties form the core foundation of reliable database systems. They ensure that even in complex, concurrent, and failure-prone environments, data remains:

  • Accurate
  • Consistent
  • Safe
  • Durable

Mastering transactions is essential for building robust applications, especially in systems where correctness is critical.

🏷️ Tags