Presentation is loading. Please wait.

Presentation is loading. Please wait.

Transactions (Chapter 10-10.3). What is it? Transaction - a logical unit of database processing Motivation - want consistent change of state in data Transactions.

Similar presentations

Presentation on theme: "Transactions (Chapter 10-10.3). What is it? Transaction - a logical unit of database processing Motivation - want consistent change of state in data Transactions."— Presentation transcript:

1 Transactions (Chapter 10-10.3)

2 What is it? Transaction - a logical unit of database processing Motivation - want consistent change of state in data Transactions developed in 1950's e.g. banking activities - allow multiple bank tellers to read and make changes Idea of transaction formalized in 1970's

3 Why? Why interleave operations? increases throughput of the system (number of transactions that can finish in any given period) see fig. 10.3 in textbook - illustrates interleaving I/Os and CPU time

4 Problems 1) Can create inconsistent result if crash in middle of transaction 2) Can have error if concurrent execution 3) Uncertainty as to when changes become permanent. Write to disk every time? Concurrency control and Recovery from failures are needed to solve these problems

5 Systems guarantees 4 properties ACID properties Atomicity - transaction is indivisible, –all or nothing, performs transaction in entirety - solves 1) Consistency - correct execution takes DB from one consistent state to another –a logical property based on some consistency rule, implied by isolation –a good test –isolation is implemented properly - solves 2) Isolation - transactions affect each other as if not concurrent –equivalent to serial schedule (serializability) - T2 -> T1 or T1 -> T2 – takes care of 2) Durability - once a transaction completes (commits), transaction is recoverable – changes are never lost due to subsequent failures - solves 3)

6 ACID properties Atomicity ensures recovery Consistency ensures programmer and DBMS enforce consistency Isolation ensures concurrency control is utilized Durability ensures recovery

7 Operations of transactions Read and Write of data items Granularity can be rows of a table or a table (Granularity can affect concurrency) Each transaction has a number assigned to it (Tn) Transaction commit statement - causes transaction to end successfully (Cn) Transaction abort (rollback) statement - all writes undone (An) System or User can specify commit and abort

8 R’s and W’s Notation: R1(X) - transaction 1 reads data item X W2(Y) - transaction 2 write to data item Y C1 - transaction 1 commits A2 - transaction 2 aborts, rollback A series of R's and W's is a schedule (history) Allow multiple users to execute simultaneously to access tables in a common DB Concurrent access - concurrency

9 Lost Update Problem Dirty write - some value of DB is incorrect T1 T2 where A=10 R1(A) A=A-10 R2(A) (A=10) W1(A) A=A+20 W2(A) The value of A is 30 when T1 and T2 are done, but it should be 20 R1(A)R2(A)W1(A)C1(A)W2(A)C2(A)

10 Dirty Read Problem Transaction updates DB item, then fails T1 T2 where A=10 R1(A) A=A-10 W1(A) R1(C) R2(A) (A=0, reads value T1 wrote) A=A+20 W2(A) A1 - T1 fails R2(B) B=B+A W2(B) When T1 is aborted, A is reset to value of 10, values for A and B written by T2 are incorrect R1(A)W1(A)R1(C)R2(A)W2(A)A1 R2(B)W2(B)C2

11 Unrepeatable Read Transaction reads data item twice, in between values change T1 T2 where A=10 R1(A) R1(B) B=B+A W1(B) R2(A) A=A+20 W2(A) R1(A) R1(C) C=C+A W1(C) When T1 first reads A it has a value of 10, then a value of 30 R1(A)R1(B)W1(B)R2(A)W2(A)C2R1(A)R1(C)W1(C)C1

12 Degrees of Isolation degree 0 - doesn't overwrite data updated (dirty data) by other transactions with degree at least 1 degree 1 - no lost updates degree 2 - no lost updates and no dirty reads degree 3 - degree 2 plus repeatable reads

13 Serializable A transaction history is serializable if it is equivalent to some serial schedule (does not mean it is a serial schedule) The important word here is some Example of two serial schedules: R1(X)W1(X)C1 R2(X)W2(X)R2(Y)W2(Y)C2 T1<< T2 R2(X)W2(X)R2(Y)W2(Y)C2 R1(X)W1(X)C1 T2 << T1

14 Equivalence Is a given schedule equivalent to some serial schedule? R2(X)W2(X)R1(X)W1(X)R2(Y)W2(Y)C1C2 How do we answer this question? What is equivalence? Need same number of operations How to define equivalence

15 Result equivalence Result equivalent if produce same final state R1(X) (X*2)W1(X)C1 R2(X)(X+Xmod10)W2(X)C2 if X=10, X=20 if X=7, X=18 R1(X)R2(X)(X+Xmod10)W2(X)C2(X*2)W1(X)C1 if X=10, X=20 if X=7, X=28 Not used

16 Conflict equivalence First define conflicting operations R and W conflict if: 1. from 2 different transactions 2. reference same data item 3. at least one is a write

17 Conflicting operations The conflicting operations are: Ri(A) << Wj(A) read followed by a write Wi(A) << Rj(A) write followed by a read Wi(A) << Wj(A) write followed by a write Ri(A) << Rj(A) don't conflict Ri(A) << Wj(B) don't conflict

18 Conflict equivalence Conflict equivalent if order of any 2 conflicting operations is the same in both schedules R1(A)W1(A)C1 R2(A)W2(A)C2 or in reverse order R1(A)<<W2(A) R2(A)<<W1(A) W1(A)<<W2(A) W2(A)<<W1(A) W1(A)<<R2(A) W2(A)<<R1(A)

19 Conflict equivalence R1(A)R2(A)W1(A)C1W2(A)C2 R1(A)<<W2(A) W1(A)<<W2(A) R2(A)<<W1(A) NOT serializable - example of lost update

20 Example Given the following two schedules - determine if each is serializable R2(X)W2(X)R1(X)W1(X)R2(Y)W2(Y)C1C2 W2(X)<<R1(X) W2(X)<<W1(X) R2(X)<<W1(X) T2<<T1 R2(X)R1(X)W2(X)R2(Y)W1(X)C1W2(Y)C2 R2(X)<<W1(X) W2(X)<<W1(X) R1(X)<<W2(X) not equivalent

21 Strategy There is a simple algorithm for determining conflict serializability of a schedule What is the algorithm?

22 Testing for Equivalence Construct a precedence graph (aka serialization graph) 1. For each Ti in S, create node Ti in graph 2. For all Rj(X) after Wi(X) create an edge Ti->Tj 3. For all Wj(X) after Ri(X) create an edge Ti->Tj 4. For all Wj(X) after Wi(X) create an edge Ti-> Tj 5. Serializable iff no cycles If a cycle in the graph, no equivalent serial history

23 Example Is this the following schedule serializable? R1(A)R2(A)W1(A)W2(A)C1C2 T1<<T2<<T1

24 Serializability If a schedule is serializable, we can say it is correct Serializable does not mean it is serial It is difficult to test for conflict serializability because the interleaving of operations is determined by the operating system scheduler Concurrency control methods don't test for it. Instead, protocols developed that guarantee a schedule is serializable.

25 Concurrency Control Techniques Protocols - set of rules that guarantee serializability 1. locking 2. timestamps 3. multiversion 4. validation or certification

Download ppt "Transactions (Chapter 10-10.3). What is it? Transaction - a logical unit of database processing Motivation - want consistent change of state in data Transactions."

Similar presentations

Ads by Google