CS377: Database Systems Concurrency Control. Li Xiong Department of Mathematics and Computer Science Emory University

Size: px

Start display at page:

Download "CS377: Database Systems Concurrency Control. Li Xiong Department of Mathematics and Computer Science Emory University"

Sherman Simmons
5 years ago
Views:

1 CS377: Database Systems Concurrency Control Li Xiong Department of Mathematics and Computer Science Emory University 1

2 Concurrent Execution of Transactions Concurrent execution of transactions is necessary for performance reasons System throughput Response time How to ensure transaction isolation while permitting concurrent execution? T1 T2 Tn DB (consistency constraints) 2

3 Concurrency Control - Outline Serializability of a schedule Enforcing serializability by locks Other types of Concurrency Control mechanisms 3

4 Schedule A schedule S of n transactions T1, T2,, Tn is an ordering of the operations of the transactions For each transaction Ti, the operations of Ti in S must appear in the same order in which they occur in Ti. Operations from other transactions Tj can be interleaved with the operations of Ti in S. A schedule represents an actual or potential execution sequence of the transactions 4

5 Example: Initial DB state: A=25, B=25 T1: Read(A) T2: Read(A) A A+100 A A 2 Write(A) Write(A) Read(B) Read(B) B B+100 B B 2 Write(B) Write(B) Constraint: A=B 5

6 Schedule A Serial Schedule T1 T2 Read(A); A A+100 Write(A); Read(B); B B+100; Write(B); Read(A);A A 2; Write(A); Read(B);B B 2; Write(B); A B

7 Schedule A Serial Schedule T1 T2 Read(A); A A+100 Write(A); Read(B); B B+100; Write(B); Read(A);A A 2; Write(A); Read(B);B B 2; Write(B); A B

8 Schedule B Serial Schedule T1 T2 Read(A);A A 2; Write(A); Read(B);B B 2; Write(B); Read(A); A A+100 Write(A); Read(B); B B+100; Write(B); A B

9 Schedule B Serial Schedule T1 T2 Read(A);A A 2; Write(A); Read(B);B B 2; Write(B); Read(A); A A+100 Write(A); Read(B); B B+100; Write(B); A B

10 Schedule C Serializable Schedule T1 T2 Read(A); A A+100 Write(A); Read(A);A A 2; Write(A); Read(B); B B+100; Write(B); Read(B);B B 2; Write(B); A B

11 Schedule C Serializable Schedule T1 T2 Read(A); A A+100 Write(A); Read(A);A A 2; Write(A); Read(B); B B+100; Write(B); Read(B);B B 2; Write(B); A B

12 Schedule D Nonserializable T1 T2 Read(A); A A+100 Write(A); Read(A);A A 2; Write(A); Read(B);B B 2; Write(B); Read(B); B B+100; Write(B); A B

13 Schedule D Nonserializable T1 T2 Read(A); A A+100 Write(A); Read(A);A A 2; Write(A); Read(B);B B 2; Write(B); Read(B); B B+100; Write(B); A B

14 Want schedules that are good regardless of initial state and transaction semantics Serializability equivalent to a serial schedule Only look at order of read and writes Shorthand representation of schedule Sc=r1(A)w1(A)r2(A)w2(A)r1(B)w1(B)r2(B)w2(B) Sd=r1(A)w1(A)r2(A)w2(A) r2(b)w2(b)r1(b)w1(b) 14

15 Serializability - Intuition Sc=r1(A)w1(A)r2(A)w2(A)r1(B)w1(B)r2(B)w2(B) Can we transform Sc into a serial schedule by swapping actions? 15

16 Serializability - Intuition Sc=r1(A)w1(A)r2(A)w2(A)r1(B)w1(B)r2(B)w2(B) Sc =r1(a)w1(a) r1(b)w1(b)r2(a)w2(a)r2(b)w2(b) T1 T2 Sc is equivalent to a serial schedule Sc (T1,T2) 16

17 Serializability - Intuition Sd=r1(A)w1(A)r2(A)w2(A) r2(b)w2(b)r1(b)w1(b)? Sd =r2(a)w2(a) r2(b)w2(b) r1(a)w1(a) r1(b)w1(b) 17

18 Serializability - Intuition Sd=r1(A)w1(A) r2(a)w2(a) r2(b)w2(b) r1(b)w1(b)? Sd =r1(a)w1(a) r1(b)w1(b) r2(a)w2(a) r2(b)w2(b) 18

19 Serializability - Intuition Sd=r1(A)w1(A)r2(A)w2(A) r2(b)w2(b)r1(b)w1(b) T2 must precede T1 in any equivalent schedule, i.e., T2 T1 Also, T1 T2 Sd cannot be rearranged into a serial schedule => Sd is not equivalent to any serial schedule => Sd is bad 19

20 Serializability Definitions Transaction: sequence of ri(x), wi(x) actions Schedule: represents chronological order in which actions are executed Serial schedule: no interleaving of actions or transactions Conflicting actions: pairs of consecutive actions such that if their order is interchanged, the behavior of at least one of the transactions can change Involve the same database element At least one is write 20

21 Serializability Definitions S1, S2 are conflict equivalent schedules if S1 can be transformed into S2 by a series of swaps on nonconflicting actions. A schedule is conflict serializable if it is conflict equivalent to some serial schedule. 21

22 Conflict-Serializable vs. View View equivalence Serializable Swapping non-conflicting actions or Swapping write operations provided if the operation Wk(Y) of Tk is the last operation to write item Y in S, then Wk(Y) of Tk must also be the last operation to write item Y in S. Conflict-serializability is a sufficient condition (not necessary) for view serializability Example: S = r1(x) w2(x) w1(x) w3(x) Schedulers in commercial systems generally use conflict-serializability 22

23 Test for Conflict-Serializability Algorithm 21.1: Constructs a precedence graph (serialization graph) - a graph with directed edges Nodes: transactions in S An edge is created from Ti to Tj if one of the operations in Ti appears before a conflicting operation in Tj The schedule is serializable if and only if the precedence graph has no cycles. What is the conflict-equivalent serial order? 23

24 Exercise S = w3(a) w2(c) r1(a) w1(b) r1(c) w2(a) r4(a) w4(d) Is S serializable? What is P(S) for S = w1(a) r2(a) r3(a) w4(a)? 24

25 Concurrency Control - Outline Serializability Enforcing Serializability Other Types of Concurrency Control Mechanisms 25

26 How to enforce serializable schedules? Option 1: run system, recording P(S); check for P(S) cycles and declare if execution was good 26

27 How to enforce serializable schedules? Option 2: prevent P(S) cycles from occurring T1 T2.. Tn Scheduler DB 27

28 Locking Two-Phase Locking Techniques Locking is an operation which secures (a) permission to Read (b) permission to Write a data item for a transaction. Example: Lock (X). Data item X is locked in behalf of the requesting transaction. Unlocking is an operation which removes these permissions from the data item. Example: Unlock (X): Data item X is made available to all other transactions. Lock and Unlock are Atomic operations. 28

29 Database Concurrency Control Two-Phase Locking Techniques: Essential components Two locks modes: (a) shared (read) (b) exclusive (write). Shared mode: shared lock (X) More than one transaction can apply share lock on X for reading its value but no write lock can be applied on X by any other transaction. Exclusive mode: Write lock (X) Only one write lock on X can exist at any time and no shared lock can be applied by any other transaction on X. Conflict matrix Read Write Read Write Y N N N 29

30 Database Concurrency Control Two-Phase Locking Techniques: Essential components Lock Manager: Managing locks on data items. Lock table: Lock manager uses it to store the identify of transaction locking a data item, the data item, lock mode and pointer to the next data item locked. One simple way to implement a lock table is through linked list. Transaction ID Data item id lock mode Ptr to next data item T1 X1 Read Next 30

31 Rule #1: Well-formed transactions Ti: li(a) pi(a) ui(a)... Lock has to precede the action and unlock has to follow the action Rule #2 Legal scheduler S =.. li(a)... ui(a)... no lj(a) No other transactions access the item during the lock 31

32 Rule #3 Two phase locking (2PL) Ti =. li(a)... ui(a)... no unlocks no locks All lock operations precede all unlock operations in a transaction growing phase and shrinking phase Theorem: Rule #1,2,3 => Conflict Serializable Schedule 32

33 # locks held by Ti Growing Phase Shrinking Phase Time 33

34 Schedule D Nonserializable T1 T2 Read(A); A A+100 Write(A); Read(A);A A 2; Write(A); Read(B);B B 2; Write(B); Read(B); B B+100; Write(B); A B

35 Schedule D possible? T1 l1(a);read(a) A A+100;Write(A); l1(b); u1(a); Read(B) B B+100;Write(B);u1(B) T2 l2(a);read(a) A Ax2;Write(A); l2(b); u2(a) Read(B) B Bx2;Write(B);u2(B) 35

36 Database Concurrency Control Dealing with Deadlock and Starvation Deadlock T 1 T 2 read_lock (Y); read_item (Y); write_lock (X); (waits for X) Deadlock (T 1 and T 2) read_lock (X); read_item (Y); write_lock (Y); (waits for Y) T1 and T2 did follow two-phase policy but they are deadlock 36

37 Deadlock Detection Timeout If transaction waits more than L sec., roll it back! Simple scheme. Problem: hard to select L Wait-for graph 37

38 Wait-For Graph Build Wait-For graph Use lock table structures Build incrementally or periodically When cycle found, rollback victim T1 T2 T5 T4 T3 T6 T7 38

Concurrency Control. Himanshu Gupta CSE 532 CC 1

Concurrency Control. Himanshu Gupta CSE 532 CC 1 Concurrency Control CSE 532 CC 1 Concurrency Control T1 T2 Tn DB CSE 532 CC 2 Overall Goal/Outline Problems that can arise in interleaving different transactions. Interleaving = Schedule. Define what are