Chapter 2

Concurrency Controlling Techniques

1
Chapter Outline
1. Databases Concurrency Control Overview
2. Purpose of Concurrency Control
3. Concurrency Control Techniques
1. Locking
2. Timestamp
3. Optimistic
4. Multiversion
5. Lock Granularity

2
Database Concurrency Control
• Transaction Processor is divided into:
– A concurrency-control manager, or scheduler, responsible for
assuring isolation of transactions
– A logging and recovery manager, responsible for the durability
of transactions.
• The scheduler (concurrency-control manager) must assure that the
individual actions of multiple transactions are executed in such an
order that the net effect is the same as if the transactions had in fact
executed one-at-a-time.

Requests from
transactions

Lock Scheduler
Table
Read and Writes

Buffers

3
1. Purpose of Concurrency Control: for performance reasons a
DBMS has to interleave the action of several transactions.
However, this has be done carefully to ensure that concurrent
execution of such transaction is equal to some serial execution
of the same transactions so as :
– To enforce Isolation (through mutual exclusion) among
conflicting transactions.
– To preserve database consistency through consistency
preserving execution of transactions.
• A typical scheduler does its work by maintaining locks on
certain pieces of the database. These locks prevent two
transactions from accessing the same piece of data at the same
time. Example:
– In concurrent execution environment if T1 conflicts with T2
over a data item A, then the existing concurrency control
decides if T1 or T2 should get the A and the other transaction
is rolled-back or waits.
4
Example:

Bank database: 3 Accounts

A = 500
Account B = 500
Balances
C = 500

Property: A + B + C = 1500

Money does not leave the system

5
 Example
Transaction T2: Transfer
Transaction T1: Transfer
100 from A to C
100 from A to B

Read (A, t) Read (A, s)

t = t - 100 s = s - 100

Write (A, t) Write (A, s)

Read (B, t) Read (C, s)

t = t + 100 s = s + 100

Write (B, t) Write (C, s)

6
 Transaction T1 Transaction T2 A B C

500 500 500

Read (A, t)
t = t - 100
Read (A, s)
s = s - 100
Write (A, s) 400 500 500

Write (A, t) 400 500 500

Read (B, t)
t = t + 100
Write (B, t) 400 600 500

Read (C, s)
s = s + 100
Write (C, s) 400 600 600
Schedule 7
400 + 600 + 600 = 1600
 Transaction T1 Transaction T2 A B C

500 500 500

Read (A, t)
t = t - 100
Write (A, t) 400 500 500
Read (A, s)
s = s - 100
Write (A, s) 300 500 500
Read (B, t)
t = t + 100
Write (B, t) 300 600 500

Read (C, s)
s = s + 100
Alternative Write (C, s) 300 600 600
Schedule
300 + 600 + 600 = 1500 8
So What ?
2. Concurrency Control Techniques
• Basic concurrency control techniques:
– Locking,
– Timestamping
– Optimistic methods
• The First two are conservative approaches: delay
transactions in case they conflict with other transactions.
• Optimistic methods assume conflict is rare and only check
for conflicts at commit.

9
 2.1. Locking
• Lock is a variable associated with a data item that describes the status of
the data item with respect to the possible operations that can be applied
to it.
• Generally, a transaction must claim a shared (read) or exclusive (write)
lock on a data item before read or write.
• Lock prevents another transaction from modifying item or even reading
it, in the case of a write lock.
– 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.
10
 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 Read Write

Y N
Write

N N

11
When we use the shared/exclusive locking scheme, the
system must enforce the following rules:
1. A transaction must issue the operation read_lock(x) or
write_lock(x) before any read_item (x) operation is
performed in T
2. A transaction T must issue the operation write_lock(x)
before any write_item (x) operation is performed in T
3. A transaction T must issue the operation unlock_lock(x)
after all read_item(x) and write_item (x) operation are
completed in T
4. A transaction will not issue a read_lock(x) operation if it
already holds a read(shared) lock or a write (exclusive)
lock on item X
5. A transaction will not issue a write_lock(x) operation if it
already holds read(shared) lock or write(exclusive) lock
on item x
6. A transaction T will not issue an unlock(x) operation
unless it already holds a read(shared) lock or a
write(exclusive) lock on item x
Slide 12
 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

 Database requires that all transactions should be well-formed.

A transaction is well-formed if:
• It must lock the data item before it reads or writes to it.
• It must not lock an already locked data items and it must
not try to unlock a free data item.
13
 Locking - Basic Rules
• It has two oprerations : Lock_item(X) and unLock_item(X)
• A transaction request access to an item X by first issuing a
lock_Item(x) opreation .
• If lock (x)=1, the transaction is forced to wait.
• If lock (X)= 0; it is set to 1 and the transaction is allowed to
access x
• When a transaction finished operation on X it issues an
Unlock _item operation which set lock(x) to 0 so that X may
be accessed by another transaction
• If transaction has shared lock on item, can read but not update
item.
• If transaction has exclusive lock on item, can both read and
update item.
• Reads cannot conflict, so more than one transaction can hold
shared locks simultaneously on same item.
• Exclusive lock gives transaction exclusive access to that item. 14
 • The following code performs the read operation: read_lock(X):
B: if LOCK (X) = “unlocked” then
begin
LOCK (X)  “read-locked”;
no_of_reads (X)  1;
end
else if LOCK (X)  “read-locked” then
no_of_reads (X)  no_of_reads (X) +1
else begin
wait (until LOCK (X) = “unlocked” and
the lock manager wakes up the transaction);
go to B
end;
• The following code performs the write lock operation: write_lock(X):
B: if LOCK (X) = “unlocked” then
LOCK (X)  “write-locked”;
else
wait (until LOCK(X) = “unlocked”
and the lock manager wakes up the transaction);
goto B
end; 15
• Lock conversion
– Lock upgrade: existing read lock to write lock
if Ti has a read-lock (X) and Tj has no read-lock (X) (i
 j) then
convert read-lock (X) to write-lock (X)
else
force Ti to wait until Tj unlocks X
– Lock downgrade: existing write lock to read lock
Ti has a write-lock (X) (*no transaction can have any lock on X*)
convert write-lock (X) to read-lock (X)
• Using such locks in the transaction do not guarantee
serializability of schedule on its own: example

16
Example : Incorrect Locking Schedule
T1 T2
Write Lock( X)
Read (X)
Schedule: X=X+100
Write(X)
Unlock(X)
Write Lock( X)
Read (X)
X= X*1.1
Write(X)
Unlock(X)
Write Lock( Y)
Read (Y)
Y= Y*1.1
Write(Y)
Unlock(Y)
Commit
write_lock(Y)
read(Y)
Y= Y-100
write(Y)
unlock(Y)
17
commit
• If at start, X = 100, Y = 400, result should be:

– X = 220, y = 330, if T1 executes before T2, or

– X = 210, Y = 340, if T2 executes before T1

• However, result gives X= 220 and Y = 340.

• S is not a serializable schedule. Why?
• Problem is that transactions release locks too soon, resulting
in loss of total isolation and atomicity.
• To guarantee serializability, we need an additional protocol
concerning the positioning of lock and unlock operations in
every transaction.

18
2.2 Two-Phase Locking Techniques: The algorithm
 Transaction follows 2PL protocol if all locking operations precede
first unlock operation in the transaction.
• Every transaction can be divided into Two Phases: Locking (Growing)
& Unlocking (Shrinking)
– Locking (Growing) Phase:
• A transaction applies locks (read or write) on desired data items
one at a time.
• acquires all locks but cannot release any locks.
– Unlocking (Shrinking) Phase:
• A transaction unlocks its locked data items one at a time.
• Releases locks but cannot acquire any new locks.
• Requirement:
– For a transaction these two phases must be mutually exclusive, that
is, during locking phase unlocking phase must not start and during
unlocking phase locking phase must not begin.

# locks
held by
Ti

Growing Phase Shrinking Phase Time 19

Example
T1 T2
Result
read_lock (Y); read_lock (X);
Initial values: X=20; Y=30
read_item (Y); read_item (X);
unlock (Y); unlock (X); Result of serial execution
write_lock (X); Write_lock (Y); T1 followed by T2
read_item (X); read_item (Y); X=50, Y=80.
X:=X+Y; Y:=X+Y;
Result of serial execution
write_item (X); write_item (Y);
unlock (X); unlock (Y); T2 followed by T1
X=70, Y=50

20
 T1 T2

read_lock (Y);

read_item (Y);
Result
unlock (Y); X=50; Y=50
Nonserializable because it
read_lock (X); violated two-phase policy.
read_item (X);

unlock (X);
Time write_lock (Y);

read_item (Y);
Y:=X+Y;
write_item (Y);
unlock (Y);
write_lock (X);
read_item (X);
X:=X+Y;
write_item (X);
unlock (X);
21
T’1 T’2
read_lock (Y); read_lock (X);
read_item (Y); read_item (X); T’1 and T’2 follow two-phase
write_lock (X); Write_lock (Y); policy but they are subject to
deadlock, which must be dealt
unlock (Y); unlock (X); . with
read_item (X); read_item (Y);
X:=X+Y; Y:=X+Y;
write_item (X); write_item (Y);
unlock (X); unlock (Y);

• If every transaction in the schedule follows the 2PL protocol ,

the schedule is guaranteed to be serializable
• Limitation
– It may limit the amount of concurrency control that can occur in
the schedule . How? (reading assignment)
22
• Remark:
– 2PL protocol guarantees serializability but it doesn’t permit
all the possible serializability schedule
– The use of this locking can cause two additional problems
• Deadlock
• starvation

23
Dealing with Deadlock and Starvation
• Deadlock
– It is a state that may result when two or more transaction are
each waiting for locks held by the other to be released
– Example :
T1 T2
read_lock (Y);
read_item (Y);
read_lock (X);
read_item (X);
write_lock (X);
write_lock (Y);
– T1 is in the waiting queue for X which is locked by T2
– T2 is on the waiting queue for Y which is locked by T1
– No transaction can continue until the other transaction
completes
– T1 and T2 did follow two-phase policy but they are deadlock
• So the DBMS must either prevent or detect and resolve such 24
deadlock situations
• There are possible solutions : Deadlock prevention, deadlock detection
and avoidance ,and lock timeouts
i. Deadlock prevention protocol: two possibilities
 The conservative two-phase locking
− Every transaction locks all data items it refers/needs to before it
begins execution.
− Limitation : It restricts concurrency
 Transaction Timestamp( TS(T) )
 We can prevent deadlocks by giving each transaction a priority
and ensuring that lower priority transactions are not allowed to
wait for higher priority transactions (or vice versa ).
 One way to assign priorities is to give each transaction a
timestamp when it starts up.
 it is a unique identifier given to each transaction based on time in
which it is started. i.e if T1 starts before T2 , TS(T1)<TS(T2)
 The lower the timestamp, the higher the transaction's priority, that
is, the oldest transaction has the highest priority.
 If a transaction Ti requests a lock and transaction Tj holds a
conflicting lock, the lock manager can use one of the
following two policies: Wait-die & Wound-wait
25
• Wait-die
– If Ti has higher priority, it is allowed to wait; otherwise it is
aborted.
– An older transaction is allowed to wait on a younger transaction.
– A younger transaction requesting an item held by an older
transaction is aborted
– If TS(Ti) < TS(Tj), then (Ti older than Tj)Ti is allowed to wait.
– Otherwise (Ti younger than Tj)Abort Ti (Ti dies) and restart it
later with the same timestamp

T1(ts =10)

wait

wait T2 (ts =20)

wait

T3 (ts =25)
26
• Wound-wait
– The opposite of wait-die
– If Ti has higher priority, abort Tj; otherwise Ti waits.
– A younger transaction is allowed to wait on an older one
– An older transaction requesting an item held by a younger
transaction preempts the younger transaction by aborting it.
– If TS(Ti) < TS(Tj), then (Ti older than Tj),Abort Tj (Ti wounds
Tj) and restart Tj later with the same timestamp
– Otherwise (Ti younger than Tj)Ti is allowed to wait

T1(ts =25)

wait

T2 (ts =20)
wait wait

T3 (ts =10) 27
  Remark:
− Both methods ended by aborting the younger of the two
transaction that may be involved in the deadlock
− Limitation:
− Both techniques may cause some transaction to be
aborted and restarted needlessly even though those
transactions may never actually cause a deadlock.
• Another group of protocols that prevent deadlock do not require
timestamps
• No waiting (NW) – If transaction is unable to obtain lock, it will be
immediately aborted and restarted after some time delay
– This method cause transactions to abort and restart transactions
needlessly
• Cautious waiting (CW) – Suppose that transaction Ti tries to lock an item X
but is not able to do so because X is locked by some other transaction Tj with a
conflicting lock; the cautious rule suggest that :
• If Tj is not blocked (not waiting for some other locked item), then Ti is blocked
and allowed to wait; otherwise abort Ti
28
ii. Deadlock Detection and resolution
– In this approach, deadlocks are allowed to happen
– The scheduler maintains a wait-for-graph for detecting cycle.
– When a chain like: Ti waits for Tj waits for Tk waits for Ti or Tj
occurs, then this creates a cycle.
– When the system is in the state of deadlock , some of the
transaction should be aborted by selected (victim) and rolled-back
– This can be done by aborting those transaction: that have made the
least work, the one with the lowest locks, and that have the least #
of abortion and so on
– Example:

29
iii.Timeouts
– It uses the period of time that several transaction have been
waiting to lock items
– It has lower overhead cost and it is simple
– If the transaction wait for a longer time than the predefined
time out period, the system assume that may be deadlocked and
aborted it
 Starvation
– Starvation occurs when a particular transaction consistently
waits or restarted and never gets a chance to proceed further
while other transaction continue normally
– This may occur , if the waiting method for item locking:
• Gave priority for some transaction over others
• Problem in Victim selection algorithm- it is possible that the
same transaction may consistently be selected as victim and
rolled-back .example In Wound-Wait
– Solution
• FIFO
• Allow for transaction that wait for a longer time
• Give higher priority for transaction that have been aborted for 30
many time
3.Timestamp based concurrency control algorithm
• Timestamp
– In lock based concurrency control , conflicting actions of different
transactions are ordered by the order in which locks are obtained.
– But here, Timestamp values are assigned based on time in which the
transaction are submitted to the system using the current date & time of
the system
– A monotonically increasing variable (integer) indicating the age of an
operation or a transaction.
– A larger timestamp value indicates a more recent event or operation.
– Timestamp based algorithm uses timestamp to serialize the execution
of concurrent transactions.
– It doesn’t use lock, thus deadlock cannot be occurred
– In the timestamp ordering, conflicting operation in the schedule
shouldn’t violate serilazable ordering
– This can be achieved by associating timestamp value (TS) to each
database item which is denoted as follow:

31
 – Time stamp ordering algorithm associates two time stamp values (TS)
with each database item X
1. Read_TS(x) : the read time stamp of x: This is the largest
timestamp among all the timestamps of transactions that have
successfully read item x
– Read_TS(X)=TS(T) where T is the youngest transaction that
has read X successfully
2. Write_TS(X) : This is the largest of all the timestamps of
transactions that have successfully written item x – that is,
write_TS(x) = TS(T), where T is the youngest transaction that
has written x successfully
• Basic Timestamp Ordering (TO):
– whenever some transaction T tries to issue a read_item(X) or a
write_item(X) operation, the basic TO algorithm compares the
timestamp of T with read_TS(X) and write_TS(X)
• The concurrency control algorithm must check whether conflicting
operations violate the timestamp ordering as in the two cases provided in
the next slide:

32
1. Transaction T issues a write_item(X) operation:
a) If read_TS(X) > TS(T) or if write_TS(X) > TS(T), then a younger
transaction has already read or written the data item so abort and
roll-back T and reject the operation.
This is because a younger transaction has already read or written the
value before T had the chance to write X;
b) If the condition in part (a) does not exist, then execute
write_item(X) of T and set write_TS(X) to TS(T)
2. Transaction T issues a read_item(X) operation:
a) If write_TS(X) > TS(T), then abort and roll-back T and reject the
operation. This is because a younger transaction has already written
the value before T had the chance to read X;
b) If write_TS(X)  TS(T), then execute read_item(X) of T and set
read_TS(X) to the larger of TS(T) and the current read_TS(X)

Strict TS Ordering
Thomas’s Write Rule

Slide 33
4. Multiversion Concurrency Control Techniques
– This approach maintains a number of versions of a data item
and allocates the right version to a read operation of a
transaction.
– Thus unlike other mechanisms a read operation in this
mechanism is never rejected.
– This algorithm uses the concept of view serilazabilty than
conflict serialiazabilty
– Side effect:
• Significantly more storage (RAM and disk) is required to
maintain multiple versions. To check unlimited growth of
versions, a garbage collection is run when some criteria is
satisfied.

34
5. Validation (Optimistic) Concurrency Control Schemes
– This technique allow transaction to proceed asynchronously
and only at the time of commit, serializability is checked &
– transactions are aborted in case of non-serializable
schedules.
– It has three phases: Read, Validation , and Write phase

i. Read phase:
– A transaction can read values of committed data items.
ii. Validation phase:.
− If the transaction Ti decides that it wants to commit, the DBMS checks whether
the transaction could possibly have conflicted with any other concurrently
executing transaction.
− While one transaction ,Ti, is being validated , no other transaction can be
allowed to commit
− This phase for Ti checks that, for each transaction Tj that is either
committed or is in its validation phase, one of the following conditions
holds:
35
  Tj completes its write phase before Ti starts its read phase.
 Ti starts its write phase after Tj completes its write phase and the
read set of Ti has no item in common with the write set of Tj
 Both the read_set and write_set of Ti have no items in common with
the write_set of Tj, and Tj completes its read phase before Ti
completes its read phase.
− When validating Ti, the first condition is checked first for each
transaction Tj, since (1) is the simplest condition to check. If (1) is
false then (2) is checked and if (2) is false then (3 ) is checked.
− If none of these conditions holds, the validation fails and Ti is aborted.
iii. Write phase:
 On a successful validation, transactions’ updates are applied to the
database; otherwise, transactions are restarted.

36
6. Multiple Granularity Locking
 A lockable unit of data defines its granularity
 Granularity can be coarse (entire database) or it can be fine (an attribute of a
relation).
 Example of data item granularity:
 A field of a database record
 A database record
 A disk block/ page
 An entire file
 The entire database
 Data item granularity significantly affects concurrency control performance.
 Thus, the degree of concurrency is low for coarse granularity and high for
fine granularity.
 Example:
 A transaction that expects to access most of the pages in a file should
probably set a lock on the entire file , rather than locking individual
pages or records
 If a transaction that requires to access relatively few pages of the file , it
is better to lock those pages
 Similarly , if a transaction access several records on a page , it should
lock the entire page and if it access just a few records , it should lock
some those records.
 This example will hold true , if a lock on the node locks that node and 37
implicitly all its descendants
 • The following diagram illustrates a hierarchy of granularity
from coarse (database) to fine (record).
Fig 1: Granularity
hierarchy

• To manage such hierarchy, in addition to read(S) and write(X), three additional locking
modes, called intention lock modes are defined:
– Intention-shared (IS): indicates that a shared lock(s) will be requested on some
descendent node(s).
– Intention-exclusive (IX): indicates that an exclusive lock(s) will be requested on
some descendent node(s).
– Shared-intention-exclusive (SIX): indicates that the current node is locked in shared
mode but an exclusive lock(s) will be requested on some descendent nodes(s).
– If a transaction needs to read an entire file and modifiy a few of the records in it.
i.e it needs S lock on a file & IX lock on some of the contained object
38
• The set of rules which must be followed for producing serializable
schedule are
1. The lock compatibility must adhered to.
2. The root of the tree must be locked first, in any mode.
3. A node N can be locked by a transaction T in S or IX mode
only if the parent node is already locked by T in either IS or IX
mode.
4. A node N can be locked by T in X, IX, or SIX mode only if the
parent of N is already locked by T in either IX or SIX mode.
5. T can lock a node only if it has not unlocked any node (to
enforce 2PL policy).
6. T can unlock a node, N, only if none of the children of N are
currently locked by T.
P
Summary :
Parent Child can be
locked in locked in
IS IS, S
IX IS, S, IX, X, SIX
S
SIX
[IS, S] not necessary
X, IX C 39
X none
  To lock a node in S mode, a transaction must first lock all its ancestors
in IS mode. Thus, if a transaction locks a node in S mode, no other
transaction can have locked any ancestor in X mode;
 Similarly, if a transaction locks a node in X mode, no other transaction
can have locked any ancestor in S or X mode.
 These two cases ensures that no other transaction holds a lock on an
ancestor that conflicts with the requested S or X lock on the node.

– These locks are applied using the following compatibility matrix:

Requestor
IS IX S SIX X
IS yes yes yes yes no
IX yes yes no no no
Holder S yes no yes no no
SIX yes no no no no
X no no no no no

• consider the following three example based Fig 1 diagram given in slide 38
• T1 wants to update record r111 and r211
• T2 wants to update all records on page p12
40
• T3 wants to read records r11j and the entire f2 file
Multiple Granularity Locking: Lock operation to show a serializable execution
T1 T2 T3
IX(db)
IX(f1)
IX(db)
IS(db)
IS(f1)
IS(p11)
IX(p11)
X(r111)
IX(f1)
X(p12)
S(r11j)
IX(f2)
IX(p21)
X(r211)
Unlock (r211)
Unlock (p21)
Unlock (f2)
S(f2)
unlock(p12)
unlock(f1)
unlock(db)
unlock(r111)
unlock(p11)
unlock(f1)
unlock(db)
unlock (r111j)
unlock (p11)
unlock (f1)
unlock(f2)
unlock(db) 41
 Quiz (6%)
1. Why binary locks are more stricter and highly conservative than
shared/exclusive locks? Explain with example *2.5pts

2. The growing and shrinking phase of 2PL policy are mutually exclusive.
Explain what does it mean? *2pts

3. Consider MGL protocol, if the parent is locked in SIX mode, mention all the
possible modes that its child can be locked in? *1.5pts

42