Timestamp-based protocols in DBMS Theory - Time & Space Complexity

Time Complexity: Timestamp-based protocols

O(n * m)

Understanding Time Complexity

We want to understand how the time needed for timestamp-based protocols changes as the number of transactions grows.

Specifically, how does the system handle more transactions and what costs increase?

Scenario Under Consideration

Analyze the time complexity of the following timestamp-based concurrency control steps.

// For each transaction T with timestamp TS(T):
// For each data item X accessed by T:
//   If TS(T) < read_TS(X), abort T
//   Else if TS(T) < write_TS(X), abort T
//   Else allow read or write and update TS(X)
    

This code checks timestamps to decide if a transaction can read or write a data item safely.

Identify Repeating Operations

Look at what repeats as input grows.

• Primary operation: Checking timestamps for each data item accessed by each transaction.
• How many times: For each transaction, the protocol checks every data item it uses.

How Execution Grows With Input

As the number of transactions and data items accessed grows, the checks increase proportionally.

Input Size (n)	Approx. Operations
10 transactions	About 10 x data items per transaction checks
100 transactions	About 100 x data items per transaction checks
1000 transactions	About 1000 x data items per transaction checks

Pattern observation: The total checks grow roughly in direct proportion to the number of transactions and their data accesses.

Final Time Complexity

Time Complexity: O(n * m)

This means the time to process transactions grows linearly with the number of transactions and the number of data items accessed per transaction.

Common Mistake

[X] Wrong: "Timestamp checks happen only once per transaction regardless of data items."

[OK] Correct: Each data item accessed requires a timestamp check, so the total work depends on all data accesses, not just the number of transactions.

Interview Connect

Understanding how timestamp-based protocols scale helps you explain concurrency control in databases clearly and confidently.

Self-Check

What if we changed the protocol to check timestamps only at the start and end of transactions? How would the time complexity change?

Practice

1. What is the main purpose of timestamps in timestamp-based protocols in DBMS?

easy

A. To give each transaction a unique order number

B. To encrypt the data during transactions

C. To store the transaction results permanently

D. To backup the database automatically

Solution

1. Step 1: Understand the role of timestamps

   Timestamps assign a unique number to each transaction to order them.

2. Step 2: Identify the main purpose in protocols

   This ordering helps the system decide which transaction can read or write data safely.

3. Final Answer:

   To give each transaction a unique order number -> Option A

4. Quick Check:

   Timestamps = Unique order number [OK]

Hint: Timestamps order transactions uniquely [OK]

Common Mistakes:

• Confusing timestamps with encryption keys
• Thinking timestamps store data permanently
• Assuming timestamps handle backups

2. Which of the following is the correct way to assign a timestamp to a transaction in a timestamp-based protocol?

easy

A. Assign the current system time as the timestamp

B. Assign a random number as the timestamp

C. Assign the transaction ID as the timestamp

D. Assign the timestamp after the transaction commits

Solution

1. Step 1: Understand timestamp assignment

   Timestamps are usually assigned based on the current system time to reflect order.

2. Step 2: Evaluate options

   Random numbers or transaction IDs do not guarantee correct ordering; timestamps must be assigned before commit.

3. Final Answer:

   Assign the current system time as the timestamp -> Option A

4. Quick Check:

   Timestamp = Current system time [OK]

Hint: Use system time for timestamps [OK]

Common Mistakes:

• Using random numbers instead of time
• Assigning timestamp after commit
• Confusing transaction ID with timestamp

3. Consider two transactions T1 and T2 with timestamps TS(T1) = 5 and TS(T2) = 10. If T1 tries to write a data item that T2 has already written, what will happen according to timestamp-based protocols?

medium

A. T1 will overwrite T2's write because it started earlier

B. T2 will be rolled back because it started later

C. T1 will be rolled back to prevent overwriting newer data

D. Both transactions will commit successfully

Solution

1. Step 1: Compare timestamps of T1 and T2

   T1 has timestamp 5 (older), T2 has timestamp 10 (newer).

2. Step 2: Apply timestamp protocol rule

   Older transactions cannot overwrite changes made by newer transactions, so T1 must be rolled back.

3. Final Answer:

   T1 will be rolled back to prevent overwriting newer data -> Option C

4. Quick Check:

   Older transaction rollback = T1 rollback [OK]

Hint: Older transactions cannot overwrite newer writes [OK]

Common Mistakes:

• Thinking older transactions can overwrite newer writes
• Assuming both transactions commit
• Confusing which transaction rolls back

4. Identify the error in this scenario: Transaction T3 with timestamp 15 reads a data item last written by T4 with timestamp 20. According to timestamp-based protocols, what is wrong if T3 reads the old value?

medium

A. No error, reading old value is allowed

B. T4 should roll back because T3 read first

C. T3 can read any value without restrictions

D. T3 should not read because it is older than T4's write

Solution

1. Step 1: Compare timestamps of T3 and T4

   T3 has timestamp 15 (older), T4 has timestamp 20 (newer).

2. Step 2: Apply read rule in timestamp protocol

   Older transactions cannot read data written by newer transactions because it causes inconsistency.

3. Final Answer:

   T3 should not read because it is older than T4's write -> Option D

4. Quick Check:

   Older transaction reading newer write = Not allowed [OK]

Hint: Older transactions cannot read newer writes [OK]

Common Mistakes:

• Assuming older transactions can read newer writes
• Thinking newer transaction rolls back
• Believing reading old value is always allowed

5. In a timestamp-based protocol, if a transaction T5 with timestamp 25 tries to write a data item that was last read by transaction T6 with timestamp 30, what should happen to maintain consistency?

hard

A. T6 should be rolled back because it read before T5's write

B. T5 should be rolled back because it is older than T6

C. Both transactions can proceed without rollback

D. T5 can write only if T6 commits first

Solution

1. Step 1: Analyze timestamps of T5 and T6

   T5 has timestamp 25 (older), T6 has timestamp 30 (newer).

2. Step 2: Apply write rule in timestamp protocol

   Older transaction T5 cannot write if a newer transaction T6 has already read the data, to avoid inconsistency.

3. Step 3: Decide action to maintain consistency

   T5 must be rolled back to prevent overwriting data read by newer transaction T6.

4. Final Answer:

   T5 should be rolled back because it is older than T6 -> Option B

5. Quick Check:

   Older write after newer read = Rollback older [OK]

Hint: Older writes after newer reads cause rollback [OK]

Common Mistakes:

• Allowing older writes after newer reads
• Rolling back newer transaction incorrectly
• Assuming both can proceed safely