CS6601-Unit 4 Distributed Systems

BY
NANDAKUMAR P,
AP/CSE,
CAHCET

INTRODUCTION
In a centralized system: all processes reside on the
same system utilize the same clock.
In a distributed system: like synchronize everyone’s
watch in the classroom.
Time is monotonous

Measuring Time
Traditionally time measured astronomically
Transit of the sun (highest point in the sky)
Solar day and solar second
Problem: Earth’s rotation is slowing down
Days get longer and longer
300 million years ago there were 400 days in the year
Modern way to measure time is atomic clock
Based on transitions in Cesium-133 atom
Still need to correct for Earth’s rotation
Result: Universal Coordinated Time (UTC)
UTC available via radio signal, telephone line, satellite (GPS)

Global Time
Global Time is utilized to provide timestamps for
processes and data.
 Physical clock: concerned with “People” time
 Logical clock: concerned with relative time and
maintain logical consistency

Physical Clocks
There are two aspects:
Obtaining an accurate value for physical time
Synchronizing the concept of physical time
throughout the distributed system
These can be implemented using centralized
algorithms or distributed algorithms

Obtaining an Accurate Physical
Time
A physical time server is needed to access the
current time from a universal time coordinator
(UTC).
Two sources for UTC:
 WWV shortwave radio station in Ft. Collins,
Colorado
 Geostationary Operational Environmental
Satellites (GEOS)

Synchronizing Physical Time
The difference in time between two clocks due to
drifting is defined as clock skew.
As long as any and every two clocks differ by a value
less than the maximum skew value, the time service is
considered to be maintaining synchronization.

Skew between computer
clocks in a distributed system

How to synchronize two clocks in A
and B?
The information necessary to read the value must
be communicated across the network to location
B.
B’s clock value must be read.
B’s clock value is communicated back to location
A.
B’s clock value is adjusted to reflect the time
necessary to travel across the network.
B’s clock value is compared to A’s clock value.

Centralized Physical Time Services
Broadcast Based
Request Driven

Broadcast Based – first approach
The centralized time server’s action:
The physical time service broadcasts periodically the current time to
members of the distributed systems.
The participants’ action:
 If a given participant’s clock is ahead of the time server’s clock, the
participant slows down its clock so that it will continually move closer
to the accurate time.
 If a participant’s clock is behind the time server’s clock, the
participant moves its clock forward. Alternatives do include gradually
speeding up the clock.

Current time
= 720
Location A
Broadcast based
Time server
Current time=740
Delay of 10
Current time=720
Adjusted current time=750
New current time=750
Location A

Current time=720
Move forward=6
Location A
Current time=740
Adjusted location A
=730
Adjusted location B
=738
Average and the new
current time=736
Time Server
Current time=732
Slow clock down to
accommodate 2
Location B
1 1
2 2
4 5
1. Current time = 740
2. My current time = 720
3. My current time = 732
4. Adjust forward = 6
5. Adjust slowdown to accommodate 2
Delay=10 Delay=6

Current time=730
Adjusted time=750
New current time=750
Current time=740
Location A Timer Server
Request for
current time
Current time=740
Delay=10

Distributed Physical Time Service
Each location broadcasts its current time at predefined set
intervals.
Once a location has broadcast its time, it starts a timer. It
then collects time messages that it receives.
 Each time message that arrives is stamped with the local
current time.
This process continues until the timer expires.
Upon the expiration of the timer, each message is adjusted
to reflect the network delay time estimated for the
message source.
At this stage, the participant calculates the average time
according to one of the following approaches:

Calculate the average of all messages
720
724
726
718
722
723
Adjusted received times

Delete the times that are above the threshold and
then average the rest.
760 X
724
726
718
702 X
723
The numbers besides X are deleted.
The rest are averaged.

Discard the highest x and the lowest x values and
then average
760 X
724
726
718
702 X
723
703 X
765 X

Logical Clocks
Why Logical Clocks?
It is difficult to utilize physical clocks to order events
uniquely in distributed systems.
The essence of logical clocks is based on the
happened-before relationship presented by Lamport.

Happen-Before Relationship
If two events, a and b, occurred at the same
process, they occurred in the order of which they
were observed. That is, a > b.
If a sends a message to b, then a > b. That is, you
cannot receive something before it is sent. This
relationship holds regardless of where events a
and b occur.
The happen-before relationship is transitive. If a happens before b and
b happens before c, then a happens before c. That is, if a > b and b > c,
then a > c.

Logical Ordering
 If T(a) is the timestamp for event a, the following relationships must hold in a
distributed system utilizing logical ordering.
If two events, a and b, occurred at the same
process, they occurred in the order of which they
were observed. That is T(a) > T(b).
If a sends a message to b, then T(a) > T(b).
If a happens before b and b happens before c, T(a)
> T(b), T(b) > T(c), and T(a) > T(c).

Process 1
Process 2
Process 3
A B
C D
E F
A>B>C>D>F E

Lamport’s Algorithm
Each process increments its clock counter
between every two consecutive events.
If a sends a message to b, then the message must
include T(a). Upon receiving a and T(a), the
receiving process must set its clock to the greater
of [T(a)+d, Current Clock]. That is, if the
recipient’s clock is behind, it must be advanced to
preserve the happen-before relationship. Usually
d=1.

Process 1
Process 2
Process 3
A(1) B(2)
C(3) D(4)
E(1) F(5)

Process 1
Process 2
Process 3
A(1.1) B(2.1)
D(4.2)
E(1.3) F(5.3)
C(3.2)
A>E>B>C>D>F

Mutual Exclusion
In single-processor systems, critical regions are
protected using semaphores, monitors, and similar
constructs.
In distributed systems, since there is no shared
memory, these methods cannot be used.

A Centralized Algorithm
Advantages: It is fair, easy to implement, and requires only three
messages per use of a critical region (request, grant, release).
Disadvantages: single point of failure.
coordinator
process Request
Grant
Enter crical
section
Exit
Release

CS6601-Unit 4 Distributed Systems

Algorithm Messages
per entry/exit
Delay
before entry
Problems
Centralized 3 2 Coordinator crash
Distributed 2(n-1) 2(n-1) Crash of any process
Token ring 1 to ∞ 0 to n-1 Lost token, process
crash

Election Algorithm
The bully algorithm
When a process notices that the coordinator is no
longer responding to requests, it initiates an
election. A process, P, holds an election as follows:
P sends an ELECTION message to all processes
with higher numbers.
If no one responds, P wins the election and
becomes coordinator.
If one of the higher-ups answers, it takes over. P’s
job is done.

7
1
2
45
36 Election
7
1
2
45
36
Ok
7
1
2
45
36
Election
7
1
2
45
36 Ok
7
1
2
45
36
Coordinator

A Ring Algorithm
1
2
4
5
3
7
6
2
2 3
2 3 42 3 4 5
2 3 4 5 6
2 3 4 5 6 1
6

Atomic Transactions
All the synchronization techniques we have studied
so far are essentially low level, like semaphores.
What we would really like is a much higher-level
abstraction such as atomic transaction.

For Example
 Atomic bank transactions:
1. Withdraw(amount, account1)
2. Deposit(amount, account2)

Stable Storage
Stable storage is designed to survive anything except
major calamities such as floods and earthquakes.
Stable storage can be implemented with a pair of
ordinary disks.
Stable storage is well suited to applications that
require a high degree of fault tolerance, such as
atomic transactions.

s a
o h
t f
b w
s a
t f
b w
s a’
o h
t f
b w
s a
t f
b w
s a
o h
t f
b w
s a
t f
b w
Stable storage
Stable storage Stable storage
Drive 1
Drive 2
(a) Stable storage (b) Crash after drive 1 is updated © Bad spot

Transaction Primitives
 1 BEGIN_TRANSACTION: Mark the start of a transaction.
 2 END_TRANSACTION: Terminate the transaction and try to commit.
 3 ABORT_TRANSACTION: Kill the transaction; restore the old values.
 4 READ: Read data from a file (or other object).
 5 WRITE: Write data to a file (or other object).
 For example,
 BEGIN_TRANSACTION
 reserve Austin-Houston;
 reserve Houston-Los Angeles;
 reserve Los Angeles-Seatle;
 END_TRANSCATION

Properties of Transactions
 Atomic: To the outside world, the transaction
happens indivisibly.
Consistent: The transaction does not violate system
invariants.
 Isolated: Concurrent transactions do not interfere
with each other.
 Durable: Once a transaction commits, the changes
are permanent.

Isolated or serializable
Isolated or serializable means that if two or more
transactions are running at the same time, to each of
them and to other processes, the final result looks as
though all transactions ran sequentially in some
(system dependent) order.

An Example
 BEGIN_TRANACATION
 X = 0;
 X=X+1;
 END_TRANSACTION
 (a)
 X=0;
 X= X+2;
 END_TRANSACTION
 (b)
 X=0;
 X=X+3;
 END_TRANSACTION
 (c )

Schedule 1 x=0; x=x+1; x=0; x=x+2; x=0; x=x+3; legal
Schedule 2 x=0; x=0; x=x+1; x=x+2; x=0; x=x+3; legal
Schedule 3 x=0; x=0; x=x+1; x=0; x=x+2; x=x+3; illegal

Nest Transactions
Transactions may contain subtransactions, often
called nested transactions.
If the subtransaction commits and the parent
transaction aborts, the permanence applies only to
top-level transactions.

Implementation
Private Workspace
1 02
0
1
2
1 02 1 2
Index
0’ 3’ 0 3
Index
Index
0
1
2
0’
1
2
3’
0
1
2
3
Private workspace

Writeahead log
x=0;
y=0;
BEGIN_TRANSACTION
x=x+1; log: x=0/;
y=y+2; log: x=0/1; y=0/2;
x=y * y; log: x=0/1; y=0/2; x=1/4;
END_TRANSACTION

Achieving atomic commit in a
distributed system
Two-Phase Commit Protocol
Coordinator Subordinates
Write “Prepare” in the log
Send “Prepare” message
Write “Ready” in the log
Send “Ready” message
Collect all replies
Phase 1
Phase 2
Write log record (if all are ready, commit; if not, abort)
Send “Commit” message
Write “Commit” in the log
Commit
Send “Finished” message

Concurrency Control
When multiple transactions are executing
simultaneously in different processes, some
mechanism is needed to keep them out of each
other’s way.
That mechanism is called a concurrency control
algorithm.

Concurrency control algorithms
 Locking
 In the simplest form, when a process needs to
read or write a file (or other object) as part of a
transaction, it first locks the file.
 Distinguishing read locks from write locks.
 The unit of locking can be an individual record
or page, a file, or a larger item.

Two-phase locking
The process first acquires all the locks it needs
during the growing phase, then releases them
during the shrinking phase.
In many systems, the shrinking phase does not
take place until the transaction has finished
running and has either committed or aborted.
This policy is called strict two-phase locking.

Time
Growing phase Shrinking phase
Lock point
Number
of locks

Optimistic Concurrency Control
 A second approach to handling multiple transactions
at the same time is optimistic concurrency control.
 The idea is simple: just go ahead and do whatever you
want to, without paying attention to what anybody else
is doing. If there is a problem, worry about it later.

Timestamps
TRD TWR T
(α) (α) (β)
TWR TRD T
(α) (α) (β)
Write
TRDT
(γ)(β)
TWRT
(γ)(β)
Do tentative
write
Abort
AbortDo tenative
write

TWRT
(γ)(β)
TTENTT
(γ)(β)
Abort
Read
T
(α) (β)
TTENT
(γ)(α)
TWR
T
(β)
TWR
Ok
Wait
Abort

CS6601-Unit 4 Distributed Systems

More Related Content

What's hot (20)

Similar to CS6601-Unit 4 Distributed Systems (20)

More from Nandakumar P (16)

Recently uploaded (20)

CS6601-Unit 4 Distributed Systems