You are on page 1of 10

COURSE TITEL: COMPUTER NETWORKING.

COURSE CODE:CSE 311

Submitted To Submitted By
MOHAMMAD EHSAN SHAHMI CHOWDHURY Name: akash Hossen
Lecturer ID: 1810020888
Department of CSE Section: 181_DB
Department of CSE
Answer To The Question NO: 1
Answer:

Reasons behind why DNS architecture is not centralized: 


Traffic Volume:
Every internet user uses DNS service every time he/she puts an address into the browser. Every
second hundreds of millions of users is browsing internet.

Maintenance:
DNS is not centralized because if it is so there will be a huge amount of load on the DNS from each
and every server. And it will be very tough to maintain such DNS at a very broad level .

That is why there is a separate individual DNS of each and every server present in the world

Single Point OF Failure:

This error means your domain name's DNS is being served by only one name server. It is a bad idea to
use a single name server for your domain because if the name server fails, your domain will be
unreachable.

distant centralized database:


A single DNS server cannot be “close to” all the querying clients. If DNS was centralized and failed
somehow the whole internet system would go down.

Answer To The Question NO: 2


ANSWER:
Connection-Oriented Demultiplexing:

If separate sockets are created for datagrams with the same destination port but different
source ip or source port addresses then this kind of demultiplexing is called connection
oriented demultiplexing.
is only TCP viable and reuqires a four-tuple, meaning in order for the datagram to know where
to be sent it needs (source IP address, source port number, destination IP address, destination
port number).

P.T.O
Figure: connection oriented demux

connection less demultiplexing:


If the same socket is used for same destination port and any kind of source ip and source port
for demultiplexing then this kind of demultiplexing is called connection less demultiplexing
requires the “IP Address” and “destination port #” weather it be TCP or UDP

Figure: connection less demux


Difference between connectionless and connection oriented demultiplexing.

In connection oriented service authentication is needed, while connectionless service


does not need any authentication. Connection oriented protocol makes a connection
and checks whether message is received or not and sends again if an error occurs,
while connectionless service protocol does not guarantees a message delivery.

Answer To The Question NO: 3

ANSWER:

RDT protocol sends packets one by one. That means the sender has to wait for the receiver for
acknowledgement after every packet is sent. This is a very slow and inefficient process. That is
way it is time consuming. It has less utilization and poor performance.

This problem can be solved by pipelining very easily. In pipelining N number of packets are sent
then the sender may wait for acknowledgement from the receiver Thus sender don’t have to
wait for receiver after every time single packet is sent. This saves a lot of time and it’s quite
efficient. So pipelining should be used instead of rdt.

2 types of pipelining protocol:

1) Go-back-N
2) Selective repeat

Difference between these two type of pipelining protocols:

Go Back-N:

1. n Go-Back-N Protocol, if the sent frame are find suspected then all the frames are re-
transmitted from the lost packet to the last packet transmitted
2. Acknowledgement type is cumulative
Selective repeat:

1. In selective Repeat protocol, only those frames are re-transmitted which are found
suspected.
2. Acknowledgement type is individual.

Answer To The Question NO: 4


ANSWER

My ID is 181002049.

So, BSI = 0001 1000 0001 0000 0000 0010 0000 0100 1001

= 000110000001000000000010000001001001

The first 32 bits of my BSI = 00011000000100000000001000000100

1st 16 bit, 0001100000010000

2nd 16 bit, 0000001000000100

The last 32-bits of your BSI=10000001000000000010000001001001

3rd 16 bit, 1000000100000000

4th 16 bit, 0010000001001001

The middle 32-bits of your BSI= 01100000010000000000100000010010

5th 16 bit, 0110000001000000

6th 16 bit, 0000100000010010

Calculation Checksum:
16 bits of Carrey Sum
1st 0001100000010000
2nd + 0000001000000100
sum 0 0001101000010100
3rd + 1000000100000000
sum 0 1001101100010100
4th + 0010000001001001
sum 0 101110110101101
th
5 + 0110000001000000
sum 1 0001101110011101
+ 1
sum 0 0001101110011110
6th 0000100000010010
sum 0010001110100000
Checksum 1101110001011111
Checksum 56415

Answer To The Question NO: 5


ANSWER:

RDT 2.1: 
.Sender sends sequences  to handle duplicate packets. 
.Use both ACK & NAK. 
.Can check whether ACK OR NAK was corrupted.
RDT 2.2:
.Same functionality as 2.1
.It's a NAK free protocol, uses ACK only.
.Duplicate ACK at sender means same thing as NAK

Although RDT 2.1 & 2.2 have same functionality RDT 2.2 is considered better as it is NAK free.
While RDT 2.1 has to deal with both ACK & NAK.

P.T.O
Figure-A: RDT 2.1 SENDER PART

Figure-B: RDT 2.1 RECEIVER PART


Figure-C: RDT 2.1 SENDER PART

Figure-D: RDT 2.2 RECEIVER PART

You might also like