Clock synchronization in distributed system

NEED??
• Time driven systems: in statically scheduled systems
activities are started at "precise" times in different points of
the distributed system.
• Time stamps: certain events or messages are associated with
a time stamp showing the actual time when they have been
produced; certain decisions in the system are based on the
"exact" time of the event or event ordering.
• Calculating the duration of activities: if such an activity starts
on one processor and ﬁnishes on another (e.g. transmitting a
message), calculating the duration needs clocks to be
synchronized.
2

 It is impossible to guarantee that physical
clocks run at the same frequency
 Lack of global time, can cause problems
 Example: UNIX make
◦ Edit output.c at a client
◦ output.o is at a server (compile at server)
◦ Client machine clock can be lagging behind the
server machine clock

When each machine has its own clock, an event
that occurred after another event may
nevertheless be assigned an earlier time.

• Every computer is equipped with CMOS clock circuit.
These are electronic devices that count oscillations
occurring in a crystal.
• Also called timer, usually a quartz crystal, oscillating at
a well defined frequency.
• Timer is associated with two registers: A Counter and
a Holding Register, counter decreasing one at each
oscillations.
• When counter reaches zero, an interrupt is generated;
this is the clock tick.
• Clock tick have a frequency of 60-100 ticks per
second.
5

• The problems:
1. Crystals cannot be tuned perfectly. Temperature and other
external factors can also inﬂuence their frequency.
Clock drift: the computer clock differs from the real time.
2. Two crystals are never identical.
Clock skew: the computer clocks on different processors of
the distributed system show different time.
6

☞ Distributed Algorithms
• There is no particular time server.
• The processors periodically reach an agreement on the
clock value by averaging the time of neighbors clock and
its local clock.
-This can be used if no UTC receiver exists (no external
synchronization is needed). Only internal
synchronization is performed.
- Processes can run on different machines and no global
clock to judge which event happens first.
7

• Cristian’s Algorithm is centralized algorithm.
 The simplest algorithm for setting time, it issues a Remote
Procedure Call to time server and obtain the time.
 A machine sends a request to time server in “d/2” seconds,
where d=max difference between a clock and UTC.
 The time server sends a reply with current UTC when
receives the request.
 The machine measures the time delay between time
server sending the message and machine receiving it.
Then it uses the measure to adjust the clock.
8

 The best estimate of message propagation time= (T0 + T1)/2
 The new time can be set to the time returned by server plus time
that elapsed since server generated the timestamp:
Tnew = Tserver + (T0 + T1)/2
9

• It is also a Centralized algorithm and its Time server is an
Active Machine.
 The server polls each machine periodically, asking it for
the time.
 When all the results are in, the master computes the
average time.
 Instead of sending the updated time back to slaves, which
would introduce further uncertainty due to network delays,
it sends each machine the offset by which its clock needs
adjustment.
 If master machine fails, any other slave could be elected
to take over.
10

a) The time daemon sends synchronization query to other
machines in group.
b) The machines sends timestamps as a response to query.
c) The Server averages the three timestamps and tells everyone
how to adjust their clock by sending offsets.
11

• NTP is an application layer protocol
• Default port number 123
• Uses standard UDP Internet transport protocol
• Adjust system clock as close to UTC as possible
over the Internet.
• Enable sufficiently frequently resynchronizations.
(scaling well on large num
• The NTP service is provided by a network of servers
located across the Internet.
• Primary servers are connected directly to a time
source. (e.g. a radio clock receiving UTC, GPS).
bers of clients and servers)
12

 Secondary servers are synchronized with primary
servers.
 The servers are connected in a logical hierarchy
called a synchronization subnet. Each level of the
synchronization subnet is called stratum.
13

 A requests time of B at its own T1
 B receives request at its T2, records
 B responds at its T3, sending values of T2 and T3
 A receives response at its T4
 Question: what is θ = TB – TA?
Synchronization in
Distributed SystemsCS-4513 D-term 2008
14
A
B
T1
T2 T3
T4

 Question: what is θ = TB – TA?
 Assume transit time is approximately the same both
ways
 Assume that B is the time server that A wants to
synchronize to
Synchronization in
15
A
B
T1
T2 T3
T4

( ) ( )2314 TTTT −−−
 A knows (T4 – T1) from its own clock
 B reports T3 and T2 in response to NTP
request
 A computes total transit time of
Synchronization in
16
A
B
T1
T2 T3
T4

( ) ( )
2
2314 TTTT −−−
( ) ( )
2
2314
3
TTTT
T
−−−
+
 One-way transit time is approximately ½ total,
i.e.,
 B’s clock at T4 reads approximately
Synchronization in
17
A
B
T1
T2 T3
T4

 Servers organized as strata
◦ Stratum 0 server adjusts itself to WWV directly
◦ Stratum 1 adjusts self to Stratum 0 servers
◦ Etc.
 Within a stratum, servers adjust with each other
CS-4513 D-term 2008
Synchronization in Distributed
Systems 18

 Assume no central time source –
 Each system maintains its own local clock .
 No total ordering of events .
 Allow to get global ordering on events.
 Assign sequence numbers to messages –
 All cooperating processes can agree on order
of event.
20

• It is used to provide a partial ordering of events
with minimal overhead.
• It is used to synchronize the logical clock.
• It follows some simple rules:
 A process increments its counter before each event in
that process i.e. Clock must tick once between every two
events.
 When a process sends a message, it includes its
timestamp with the message.
 On receiving a message, the receiver process sets its
counter to be the maximum of the message counter and
increments its own counter .
21

 a ‘Happened Before’ b : a→b
1. If a and b are events in the same process, and a
comes before b, then a → b.
2. If
a :message sent
b : receipt of the same message
then a → b.
1. Transitive: If a → b and b → c then a → c.
2. Two distinct events a and b are said to be concurrent
if a -/->b and b -/->a
22

Figure 6-9. (a) Three processes, each with its own clock.
The clocks run at different rates.

 For two events a and b in same process p1
 C(b)= C(a)+1
 If a is sending process and b is receiving process
of pi and pj then,
 Cj(b)=max((Ci(a)+1),Cj(b))
24

Lamport’s algorithm corrects the clocks.

Clock synchronization in distributed system

More Related Content

What's hot (20)

Viewers also liked (19)

Similar to Clock synchronization in distributed system (20)

More from Sunita Sahu (7)

Recently uploaded (20)

Clock synchronization in distributed system