Transcript 04CM20145Lecture13

Dr Alwyn Barry
Dr Joanna Bryson
CM20145
Transactions
Lecture Plan
1. Basic Concepts
2. Data, Information & Knowledge
3. Data Models (The E-R Model)
4. The Relational Algebra
5. Introduction to SQL
6. Further SQL (Joins, RA Equivalences)
7. Database Design
8. Further DB Design – Normalisation
9. Architectures and Implementations
10. Integrity and Security
Lecture Plan
11. Ethics and Professional Conduct
12. Legal Issues
13. Transactions
14. Recovery
15. Concurrency Control
16. Storage and File Structure
17. Indexing and Hashing
18. Query Processing & Optimisation
January… Review, Object Relational Bridges?
A While Ago…
 Architectures and Implementations


Introductions to Transactions & Storage
Architecture concerns:
 Speed, Cost, Reliability, Maintainability.

Architectural Types:
 Centralized, Client/Server, Parallel, Distributed
 Integrity and Security


Domain Constraints
Referential Integrity
 Foreign Keys, Cascading Actions


Assertions
Triggers

Authorization
 Grant, Revoke, Roles, Audit Trails
Now: Transactions, Concurrency & Recovery
Overview
 Transaction Concepts


ACID
Possible States



Conflict
View
Others



Precedence Graphs
Conflict
View
 Schedules
 Serializability
 Testing for Serializability
 Concurrency & Recovery
Introduction to Transactions
 A transaction is a unit of program execution
that accesses and possibly updates various data
items.
 A transaction starts with a consistent database.
 During transaction execution the database may
be inconsistent.
 When the transaction is committed, the
database must be consistent.
 Two main issues to deal with:
 Failures, e.g. hardware failures and system
crashes.
 Concurrency, for simultaneous execution of
multiple transactions.
©Silberschatz, Korth and Sudarshan
Modifications & additions by S Bird, J Bryson
The ACID Test
To preserve integrity of data, the
database system must ensure:
 Atomicity: Either all operations of the
transaction are properly reflected in the
database or none are.
 Consistency: Execution of a transaction in
isolation preserves the consistency of the
database.
 Isolation: Although multiple transactions
may execute concurrently, each transaction
must be unaware of other concurrently
executing transactions; intermediate
transaction results must be hidden from other
concurrently executed transactions.
 Durability: After a transaction completes
successfully, the changes it has made to the
database persist, even if the system fails.
Example: A Fund Transfer
Transfer $50 from account A to B:  Durability: once the
1. read(A)
user notified that the
transaction complete,
2. A := A – 50
the updates to the
3. write(A)
database by the
4. read(B)
transaction must
5. B := B + 50
persist despite
6. write(B)
failures.
 Isolation: between
 Consistency: the sum of A
steps 3-6, no other
and B is unchanged by the
transaction should
execution of the transaction.
access the partially
 Atomicity: if the transaction
updated database, or
fails after step 3 and before
it would see an
step 6, the system must
inconsistent state
ensure that no updates are
(A + B will be less
reflected in the database, else
than it should be).
an inconsistency will result.
Transaction States
 Active, the initial state; the transaction stays in this
state while it is executing
 Partially committed, after the final statement has
been executed.
 Committed, after successful completion.
 Failed, after the
discovery that normal
execution can no
longer proceed.
 Aborted, after the
transaction has been
rolled back and the
database restored to
its state prior to the
start of the
transaction.
Transaction Definition in SQL
 Data manipulation languages must
include a construct for specifying
the set of actions that comprise a
transaction.
 In SQL, a transaction begins
implicitly.
 A transaction can be explicitly
ended by:
 Commit work: commits current
transaction and begins a new one.
 Rollback work: causes current
transaction to abort.
Overview
 Transaction Concepts


ACID
Possible States



Conflict
View
Others



Precedence Graphs
Conflict
View
 Schedules
 Serializability
 Testing for Serializability
 Concurrency & Recovery
Schedules & Concurrency
 Advantages to Concurrent execution (executing
transactions simultaneously):
 Increased processor and disk utilization; better throughput.
 One transaction uses CPU while another uses disk.
 Reduced average response time: short transactions need
not wait behind long ones.
 Concurrency control schemes:
 Mechanisms to achieve isolation.
 Control concurrent transactions’ interaction in order to
prevent them from destroying database consistency.
 Schedules:
 Sequences that indicate the chronological order in which
instructions of concurrent transactions are executed.
 A schedule for a set of transactions must consist of all
instructions of those transactions.
 Must preserve the order in which the instructions appear in
each individual transaction.
Example: Serial Schedule
 Let T1 transfer $50
from A to B, and T2
transfer 10% of the
balance from A to B.
 This is a serial
schedule, in which
T1 is followed by T2.
Example: Concurrent Schedule
 Let T1 and T2 be
the transactions
defined previously.
 This schedule is
not a serial
schedule, but it is
equivalent to the
previous schedule.
In both this and the sequential schedule,
the sum A + B is preserved.
Concurrency Gone Bad
 This concurrent
schedule does not
preserve the value
of A + B.
Overview
 Transaction Concepts


ACID
Possible States



Conflict
View
Others



Precedence Graphs
Conflict
View
 Schedules
 Serializability
 Testing for Serializability
 Concurrency & Recovery
Serializability
 Basic Assumption: Each transaction,
on its own, preserves database
consistency.
 That is, serial execution of transactions
preserves database consistency.
 A (possibly concurrent) schedule is
serializable if it is equivalent to a
serial schedule.
 Different forms of equivalence lead to
different kinds of serializability:
conflict and view.
 Serialization makes recovery easier,
but can slow down throughput.
Conflict Serializability
 Instructions li and lj of transactions Ti and Tj
respectively, conflict iff there exists some
item Q accessed by both li and lj, and at least
one of these instructions wrote Q.
1.
2.
3.
4.
li
li
li
li
=
=
=
=
read(Q), lj = read(Q).
read(Q), lj = write(Q).
write(Q), lj = read(Q).
write(Q), lj = write(Q).
li and lj don’t conflict.
They conflict.
They conflict
They conflict
 Intuitively, a conflict between li and lj forces
a (logical) temporal order between them.
 If li and lj are consecutive in a schedule and
they do not conflict, their results would
remain the same even if they had been
interchanged in the ordering.
Conflict Serializability (2)
 If a schedule S can be transformed into a
schedule S´ by a series of swaps of nonconflicting instructions, we say that S and S´
are conflict equivalent.
 We say that a schedule S is conflict
serializable if it is conflict equivalent to a
serial schedule.
 Example of a schedule that is not conflict
serializable:
T3
T4
read(Q)
write(Q)
write(Q)
We are unable to swap instructions in the above
schedule to obtain either the serial schedule < T3, T4 >,
or the serial schedule < T4, T3 >.
Conflict Serializability (3)
 The first example
concurrent schedule
can be transformed
into the serial one
(where T2 followed T1)
by a series of swaps of
non-conflicting
instructions.
 Therefore our
concurrent schedule is
conflict serializable.
View Serializability
 Let S and S´ be two schedules with the same
set of transactions. S and S´ are view
equivalent if the following three conditions
are met, where Q is a data item and Ti is a
transaction:
1. If Ti reads the initial value of Q in schedule S, then Ti
in schedule S´ must also read the initial value of Q.
2. If Ti executes read(Q) in schedule S, and that value
was produced by transaction Tj (if any), then
transaction Ti must in schedule S´ also read the
value of Q that was produced by transaction Tj
3. The transaction (if any) that performs the final
write(Q) operation in schedule S (for any data item
Q) must perform the final write(Q) operation in
schedule S´
NB. View equivalence is also based purely on
reads and writes
View Serializability (2)
 A schedule S is view serializable if it is
view equivalent to a serial schedule.
 Every conflict serializable schedule is
also view serializable.
 Some schedules are view-serializable
but not conflict serializable (see below).
 Every view serializable schedule that is
not conflict serializable has blind writes.
Other Notions of Serializability
 This schedule
produces the same
outcome as the
serial schedule
< T1, T5 >
 However it is not
conflict equivalent
or view equivalent
to it.
 Determining such
equivalence requires
analysis of
operations other
than read and write.
This is hard (computationally).
Overview
 Transaction Concepts


ACID
Possible States



Conflict
View
Others



Precedence Graphs
Conflict
View
 Schedules
 Serializability
 Testing for Serializability
 Concurrency & Recovery
Testing for Serializability
 Consider some schedule of a set of
transactions T1, T2, ..., Tn
 Precedence graph: a directed graph where
the vertices are transaction names.
 We draw an arc from Ti to Tj if the two
transactions conflict, and Ti accessed the data
item before Tj
 We may label the arc by the item that was
accessed.
 Example:
x
y
Schedule & Precedence Graph
T1
T2
read(X)
T3
T4
T5
read(Y)
read(Z)
read(V)
read(W)
read(W)
T
T
1
2
read(Y)
write(Y)
write(Z)
T
read(U)
read(Y)
write(Y)
read(Z)
write(Z)
read(U)
write(U)
3
T
4
Testing Conflict Serializability
 A schedule is conflict
serializable if and only if its
precedence graph is acyclic.
 Cycle-detection algorithms exist
which take order n2 time, where
n is the number of vertices in
the graph.
 If precedence graph is acyclic,
the serializability order can be
obtained by a topological
sorting of the graph. This is a
linear order consistent with the
partial order of the graph.
 For example, a serializability
order for this graph is:
T1  T2  T 4  T3  T5
Example of an acyclic
precedence graph
Testing View Serializability
 The precedence graph test for conflict
serializability must be modified to apply to a
test for view serializability.
 The problem of checking if a schedule is view
serializable is NP-complete. Thus existence of
an efficient algorithm is unlikely.
 However practical algorithms that just check
some sufficient conditions for view
serializability can still be used.
Overview
 Transaction Concepts


ACID
Possible States



Conflict
View
Others



Precedence Graphs
Conflict
View
 Schedules
 Serializability
 Testing for Serializability
 Concurrency & Recovery
Concurrency & Serializability
 Goal – to develop concurrency control
protocols that will ensure serializability.
 These protocols will impose a discipline
that avoids nonseralizable schedules.
 A common concurrency control protocol
uses locks.
 While one transaction is accessing a data
item, no other transaction can modify it.
 Require a transaction to lock the item before
accessing it.
 Topic of Lecture 15!
Recoverability
 How do we address failures
when we are running
concurrent transactions?
 Recoverable schedule: if a transaction Tj
reads a data item previously written by a
transaction Ti , the commit operation of Ti
appears before the commit operation of Tj
 This schedule is not recoverable if T9 commits
immediately after the read.
 If T8 should abort, T9 would have read (and
possibly shown to the user) an inconsistent
database state.
 A Database must ensure that schedules are
recoverable!
Summary
 Transaction Concepts






ACID
Possible States
Schedules
Serializability



Conflict
View
Others
Testing for Serializability



Precedence Graphs
Conflict
View
Concurrency & Recovery
 Next:
Recovery
Reading & Exercises
 Reading
 Silberschatz Ch: 15.
 Connolly & Begg 20.1 – 20.2.2 (very
clear!)
 Exercises:
 Silberschatz 15.1, 15.5-9.
 Connolly & Begg 20.1-3, 20.18-19