739869
CONCURRENCY
CONTROL
- Problem Definition
- Objective
- Maximize transaction
throughput(work perform) while
preventing interference among
multiple users
- Throughput: number of
transactions processed per
unit time eg >1million
transaction per second
- Maximize transaction
throughput(work perform) while
preventing interference among
multiple users
- Constraint
- No interference:
result same as
serial/sequential
effect
- Interference occurs on
commonly manipulated
data known as hot spots
- No interference:
result same as
serial/sequential
effect
- Objective
- Concurrency
Control Problem
- Lost Update Problem
e.g:slide no.11&12
- Preventing Lost Update
Problem Using 2PL
e.g:slide no.28
- Preventing Lost Update
Problem Using 2PL
e.g:slide no.28
- Uncommitted
Dependency Problem
e.g:slide no.13&14
- Preventing
Uncommitted
Using 2PL
e.g:slide no.29
- Preventing
Uncommitted
Using 2PL
e.g:slide no.29
- Lost Update Problem
e.g:slide no.11&12
- Concurrency
Control Tools
- Locking
- Obtain lock before
accessing an item
- Wait if a conflicting lock
is held
- Shared lock (S lock):
conflicts with exclusive
locks
- Exclusive lock (X lock):
conflicts with all other
kinds of locks
- Shared lock (S lock):
conflicts with exclusive
locks
- Locking Conflict
e.g:slide no.22
- Locking Granularity
e.g:slide no 23
- Obtain lock before
accessing an item
- Deadlock
- Deadlock(Mutual
Waiting)
e.g:slide no.24
- Deadlock Resolution
- Detection
- Overhead is reasonable
for deadlocks among 2 or
3 transactions
- Overhead is reasonable
for deadlocks among 2 or
3 transactions
- Timeout
- Waiting limit
- Can abort
transactions that
are not deadlocked
- Timeout interval is
difficult to determine
- Waiting limit
- Detection
- Deadlock(Mutual
Waiting)
e.g:slide no.24
- Locking
Want to create your own Mind Maps for free with GoConqr? Learn more.