Assume propagation delay is less than transmission delay on a link connecting host A to to B. If host A starts transmission at t=0, at time equal equ al to transmission delay, where is the first bit of the packet? At host A In the link At host B
uppose ! packets arri"e simultaneously at a router e"ery # seconds. uppose the packets are of si$e bits and the transmission rate of the router is % bps.
&hat should the minimum minimum "alue of # be such that the queuing delay does not grow without bound? !%' (!)*+'% (!)+'% NS/R Answer Description For the queuing delay to not increase without bound, its important that the queue is empty before the next batch of packets arrive. The time it takes to send out (transmit) packets is S/R. So T is N !"#.
0 ' - points points 0 ' attempts At this "alue of #, what is the a"erage queuing delay eperienced b y a packet? The delay experienced experienced by the the $rst packet packet in the batch is %. &y the second second packet packet is !"#. The third is '!"# and so on. veraging it yields % * !"# * '!"# * ............ * (+)!"# - " (+)! " ('#)
!'% !'% (N-1)S/2R (!)+'%
•
&it /ength
% " ' points % " ' attempts •
0n a 1bps link, where light travels at '2%34m"s, what would the bit length be in meters5 Format to use 6 (no decimals, eg. 789)
o choice
/ile transfer time &hat is is the time required in sec (from *st bit sent to to last bit recei"ed+ to send *0 1B of data?
0 ' * points points 0 ' * attempts Assume a * 2bps link with a propagation p ropagation delay of *0 ms. pecify time in sec. /ormat 3. (decimal places. 4g. 55.555, 5.555+
0 ' points points 0 ' * attempts Assume a *00 kbps link with a propagation delay of *0 ms. pecify time in sec. /ormat 3. (decimal places. 4g. 55.555, 5.555+ Hint #otal #otal time = # time 6 7ropagation time
•
&its
% " 7 points % " ' attempts !uppose two hosts and & are connected by a 1bps link of length % km. !uppose the speed of light over the link is ' 2 %34 m"s. :f a 9 1& $le were to be transferred between the hosts as back+to+back packets, how many bits will be in the link at any given time5 Format Format 6 (e.g. %%) •
/ile fragmentation A)))))%)))))B 8osts A and B are connected "ia a router %. uppose the Bandwidth of the A to % link as well as % to B link is *0 1bps. Assume propagation delay on each link is *0 us and processing speed at % is -0 us. Assume the router needs to retrie"e the entire packet before processing. 9alculate the time (in us+ it takes to transmit *0k bits from A to B.
0 ' : points 0 ' attempts Assume the *0k bits are sent as two ;k bit packets back)to back.pecify the time in us (microsec+. /ormat 3 (e.g. *000+
Hint 7ipelining plays a role here 0 ' points 0 ' attempts
Assume the *0k bits are sent as a single packet. pecify the time in us (microsec+. /ormat 3 (e.g. *000+ !ee Figure. !uppose the :"0 bus within a host has a speed of %% 1bps, would it be possible to sustain a rate of %% 1bps on the outgoing link5 ;hy or why not.
ee /igure. uppose a workstation is configured as a router with an internal I'< bus speed of *0 2bps. Assume 1A mode of memory access and that the 97> has to eamine the headers (all at once+ to determine the route of the packet. If the total headers si$e is *0 of the total packet si$e and the router has to output packets at a rate of * million packets per sec, what should be the packet si$e in bits such that the router can sustain this outgoing rate? p @ ;000 p ;000 p @ :C p :C
*0dB attenuation means the signal strength hal"ed doubled reduced by factor of ten increased by factor of ten
:f a cable su=ers 7d& attenuation per %%m and if the transmit power is %%mw. ;hat is the received power if the cable length is km5 >xpress the received power in mw. Format 6.x (>.g. .) o choice ?etermine the distance between a transmitter and receiver of a wireless @%%1hA link if the transmit power is watt, received power is %. nano watt. ssume free space path loss model. :gnore antenna gains (not covered in class). >xpress the distance in km. Format 6.xx (>.g. 7.B9) o choice
&hich among the following does !<# causes distortion of a signal? reflection mobility attenuation linear amplification
noiseless B CDA channel can support which of the following data rates (in kbps) if binary pulses are employed5 Top of Form B 8 (wrong) 8 all of the above &ottom of Form
!# #euired ++To achieve 9% kbps data rate on a 9 khA channel, what !# (in d&) is needed5 Format 6 (>.g. 7) 30
1a data rate )) A transmission line has a bandwidth of * 1h$. &hen *0 dBm power is transmitted, the recei"er recei"es )C0 dBm. #he noise le"el at the recei"er is )*00 dBm. &hat is the maimum theoretical rate (in 1bps+ that can be achie"ed on this link? /ormat 3. (4.g ;.*+ !ote dBm is nothing but power ratio in decibels (dB+ of the obser"ed power referenced to one milliwatt (m&+. o it boil downs to *0logD*0(7 in mw+. 4g. 0dBm = *0 logD*0(*mw+
Eacket error rate
% " points % " attempts
;hat is the Eacket >rror #ate (E>#) in percentage at a bit error rate of %3+ if packet siAe is %k bits.
1-(1-10 power-7)1000 Top of Form G %.G 0.01% %.%%G &ottom of Form
!ignal bandwidth
% " ' points % " ' attempts
!ee $gure. ;hat is the bandwidth in 1hA occupied by the signal if T is us5 Format 6 (e.g. ') or answer HinfH for in$nity. inf
Ilock !ynchroniAation
% " points % " attempts
;hich of the following encodings do not provide proper clock synchroniAation5 Top of Form #J NRZI 1anchester B&"9& &ottom of Form %2 ;hich of the following encodings have problem with long sequence of Ks Top of Form #J #J: 1anchester B&"9 &ottom of Form
1anchester encoding can be thought of as 6&"L& encoding. ;hat are 6,L5 Format 6,L (>.g. ,7 )B,9 o choice
Assume a bit duration of *ms. uppose the sender clock is rEunning ; slower than the recei"er clock and the recei"er samples the bit in what it belie"es to be the middle of the bit. If !%F encoding is used, at what bit position starting from *, can the recei"er possibly make an error. #his means the sender sends first bit from 0 to *.0;ms, secon d bit from *.0; to .*ms and so on. #he recei"er samples first bit at 0.;, second bit at *.; and so on. 4press answer as a bit position (starting from *+. /ormat 3 (e.g. +
B&"9&
% " points % " attempts
;hen employing B&"9& encoding, what would the below bit sequence translate to5
%% %%% %%%%%% FormatM bit sequence without spaces (>.g. %%%%%%%%) % " % points % " attempts Ian you design an encoding scheme with an even stronger requirement of no more than ' consecutive %Ks in the encoded bit stream5
;hat would be the eNciency of such a scheme5 B&"9&
IIIII
?// O F#1> Ian errors in one frame a=ect the next frame5 Top of Form yes no &ottom of Form
&yte !tuffing
% " points % " attempts
!uppose the last two bytes of the &i!LI protocol are ?/> >T6, what sequence of bytes precede I#I5 Top of Form ?/> ?/> ?/> >T6 >T6 ?/> ?/> >T6 ?/> ?/> >T6 >T6 DLE DLE DLE ETX &ottom of Form
&it !tuffing
% " points % " attempts
:f the following bit pattern is received as part of D?/I data, what does the D?/I decoder output after removing stu=ed bits5
%%%% >xpress the answer as a sequence of bits with no spaces (e.g. %%%%)
>scape characters
% " points % " attempts
Ian this sequence K ?/> ?/> >T6K appear in the data portion of the transmission when using the &isync protocol, where is the !I:: character KK. Top of Form yes no &ottom of Form >rror Iontrol
8amming distance
0 ' * points 0 ' * attempts
If a code has a hamming distance of , how many errors can it detect? /ormat 3 (e.g. +
0 ' * points 0 ' * attempts If a code has a hamming distance of , how many errors can it correct? /ormat 3 (e.g. *+ 3 :f a code contains the following code wordsM
c %%%%, c' , c7 %%%, cB %%%%%%% and c9 %%%%
;hat is the hamming distance of this code5
1r. 4rror 7rone wants to decide on whether to use error correction or detection o n a link with packet error rate of *0G);. uppose the packet si$e is 5- bits and he is considering B98 ( Bose) 9haudhuri)8ocquenghem+ code for error correction, where *0- code bits are sent for 5- data bits and 9%9 for error detection ,where - bits are added to 5- data bits.
0 ' points 0 ' attempts &hat is the transmission o"erhead of the error correction in ? #ransmission o"erhead is the percentage of HadditionalH bits transmitted per data bit. /ormat 3. (e.g. *;.-+
0 ' : points 0 ' attempts &hat is the transmission o"erhead of the error detection in ? #ransmission o"erhead is the percentage of HadditionalH bits transmitted per data bit. /ormat 3. (e.g. *;.-+
7u$$le
0 ' 0 points 0 ' * attempts A bit challenging but fun nonetheless...
An absent minded professorHs glasses are inside one of three rooms. 8e wants to send fi"e of his
students in groups to the three rooms to check for his glasses.
8ow should the professor di"ide the students into groups so that he can determine where his glasses are?
Also, how can he determine in which room his glasses are?
9ould the professor determine where his glasses are if there were on ly four students, and only one of them might lie?
1ap this problem to error correcting codes and sol"e. •
•
•
?etection
•
!ingle bit parity
% " 7 points % " attempts •
!marty Eants wants to enhance single bit party scheme and decided to use two redundant bits. 0ne bit represents the parity of all odd+numbered bits and second bit the parity of all even+numbered bits. >xampleM % is encoded as %% (assuming Keven+parityK).
;hats the hamming distance of this code5 ote message bits can be any number, redundant bits are '. nswer Format 6 (e.g. 9)
Das he done better than single bit parity in terms of hamming distance5 ?o not answer this in the box, Pust for your thinking. IIIIIIIIIIIIIIIcc Two dimensional parity % " % points % " attempts •
;hat is the general set of circumstances under which a B+bit error cannot be detected by a two dimensional parity5
Hint #ectangle IIIIIIIIIIIIIc
:nternet Ihecksum % " % points % " attempts •
The :E protocol implements :nternet checksum over Pust the :E header. s the packet passes through the routers, one $eld called TimeTo/ive (TT/) (4+bits long) in the :E header is decremented at each router. !o, each router needs to decrement this $eld and then update the checksum in the header. :s there a way to update the checksum without having to recalculate the checksum over the entire header5
Hint 8+bit word in checksum calculation is made up of two bytes. Ionsider two cases, where TT/ $eld falls in low byte order (least signi$cant byte) and another where it falls in higher byte order (most signi$cant byte). IIIIIIIIIIIIIc I#I % " ' points % " attempts •
For a generator polynomial of x37*x3'*, what is the I#I value if the message is %%%%5 >xpress answer as a bit sequence with no spaces (>g. %) I#I &urst >rrors % " % points % " attempts •
!how that any bursts of length Q k can be detected if I(x) includes a constant term (x3% term).
#?T Eacket sequence
% " 7 points % " ' attempts •
Tracing revealed the following sequence of packets sent by a sender employing stop and wait protocol. K%K and KK refer to the sequence number in the packet.
%%%%%%%%%
;hich of the following is most likely to produce the pattern. &ad forward channel (data path) &ad reverse channel (ack path) &ad timeout setting &ug in error detection algorithm (reproducible bug)
!top and wait % " % points % " attempts •
:n stop and wait protocol, suppose both sender and receiver retransmit the previous packet on receipt of a duplicate IC or data frame.
?raw a timeline to show the sequence of events which will happen if somehow the $rst data frame is duplicated. IIIIIIIIIIIIc !top and wait data rateM !uppose the stop and wait protocol is employed over a link to &, where the to & link bandwidth is 1bps with a propagation delay of '%ms and & to link bandwidth is %1bps with a propagation delay of %ms (reverse link di=erent from forward link). !uppose that the data packet siAe is %%% byte and IC siAe is %% byte. % " 7 points % " ' attempts •
;hat is the throughput achieved by the protocol5 :gnore other delays. >xpress answer in kbps. Format 6 (e.g. 9')
% " ' points % " attempts •
:f you are designing a new #?T protocol, what should your target throughput be i.e what is the maximum throughput that can (potentially) be achieved over this link5 >xpress the answer in kbps. Format 6 (>g. '9%).
!i"ing #in"ow !ending ;indowM ssume a link with data rate of 1bps and one way latency of B%ms. ssume a packet siAe of %%% bytes. % " ' points % " ' attempts •
;hat should the sending window siAe be (in packets) over this link5 Format 6 (e.g. )
% " ' points % " ' attempts •
For !;! as obtained in previous question, if #;! was set to , what would the minimum number of bits needed to represent the sequence number5 Format 6 (e.g. 4)
lternating bit protocol % " points % " attempts •
:s the alternating bit protocol the same as Ro+&ack+ protocol with a sender and receiver window siAe of 5 yes no !liding ;indow IC % " % points % " attempts •
:n the sliding window protocol, can the sender receive an IC for a packet that falls outside its current window5 :f so, specify the scenario under which this occurs as well. 1inimum sequence number
% " % points % " attempts •
?erive a general rule for the minimum sequence number space in terms of !;! and #;!.
!liding ;indow Timeline+ % " % points % " attempts •
?raw a timeline of the sliding window protocol with !;!#;!7. ssume frames B,9,8 are lost ($rst copies only). Sse a time+out of '2#TT and that the sender re+ transmits the lost frames on timeout. ?raw the timeline till the frames B,9,8 are successfully received at the receiver. !liding ;indow Timeline+' % " % points % " attempts •
!ame problem as before except that the sender retransmits frames on receiving $rst duplicate IC. ?raw a timeline of the sliding window protocol with !;!#;!7. ssume frames B,9,8 are lost ($rst copies only). Sse a time+out of '2#TT. ?raw the timeline till the frames B,9,8 are successfully received at the receiver. !liding ;indow !elective ck % " % points % " attempts •
?raw a timeline of the sliding window protocol that employs selective acknowledgment with !;!#;!B. ssume frames B,9 are lost. Sse a time+out of '2#TT. !how the timeline till frame is sent.
>thernet Dub vs switch % " % points % " attempts •
switch can connect %1bps and a %%1bps link, while a hub cannot. ;hy5 >thernet adaptor
% " points % " attempts •
Dow many >thernet adaptors can potentially be manufactured5 >xpress it as a power of '. Format '36 (e.g. '3%). >thernet ddress % " points % " attempts •
:f the $rst bit in an >thernet 1I address is a Aero, the address is what type of address5 unicast broadcast multicast Iollision domains % " ' points % " ' attempts •
:n the $gure shown, how many collision domains are there5 Format 6 (e.g. 4)
Iompeting nodes % " ' points % " attempts •
!ee Fig. :dentify the collision domains. :n the largest collision domain, how many nodes are competing for the channel5 Format 6 (e.g. @)
!witching M o need Time of TransferM !uppose Dost is connected to another Dost & via a switch. The data rate of the link connecting the hosts to the switch is #bps. ssume a packet length of / bits. :n the calculation below, ignore processing, propagation and queuing delay.
++++ ! ++++ & % " points % " attempts •
:f circuit switching was used to forward the packet, how much time would it take to send the packet from to & after connection establishment5 ssume a dedicated wire is allocated to the connection (not T?1). /"# #"/ '/"# 7/"# % " points % " attempts •
:f using packet switching, how much time would it take to send the packet from to &5 /"# '/"# 7/"# B/"#
Eayload !iAe % " B points % " ' attempts •
!uppose that %%k&ytes of user data has to be sent from host to host & passing through B switches employing packet switching. !uppose the data rate on each of the link is 1bps and that each packet carries a header of B%bytes. ;hat value of payload siAe (payload in the packet excluding header) minimiAes the total delay in delivering the user data5 Total delayM $rst bit was sent from to when last bit was received at &. :gnore propagation"processing delay. >xpress the answer in bytes. Format 6 (e.g. '%%)
Eacket vs Iircuit !witchingM :n this question, we will compare circuit with packet switching. !uppose (& for &ytes) of user data has to be sent from host to host & passing through 7 switches employing packet switching. !uppose the data rate on each of the link is 1bps and propagation delay is ms. For circuit switching, assume that a %%%& packet has to be sent from source to destination and back from destination to source before the user data can be sent. ?uring this connection set+up time, assume that each switch introduces ms of processing delay (packet fully removed from link before processing). fter connection set up, assume a dedicated wire is allocated to the connection. There is no processing delay at the switches during the data transfer phase. For packet switching, assume a payload siAe of %%%& and a header siAe of '%&. For forwarding the packet, each switch incurs ms processing delay. lso assume that the $le siAe is a multiple of %%%&. % " 9 points % " ' attempts •
For what value of is the total delay in delivering the user data ($rst bit from sender to last bit received at receiver) less for circuit switching compared to packet switching. >xpress in Cilobytes in the answer. Format 6 (>.g. '%%)
% " 7 points % " ' attempts •
For what value of (minimum), the total number of bytes sent (including payload, headers, connection+setup packets, bytes sent by sender"receiver ) in the network HlessH for circuit switching compared to packet switching5 >xpress in Cilobytes in the answer. Format 6. >.g. '%% (to indicate '%%C&)
!haring % " ' points % " ' attempts •
!uppose, you have a 1bps link being shared by 7 users at a switch. >ach user is active only %G of the time and when active needs B%%kbps .
:f packet switching was being used, for what fraction of the time (expressed in percentage) can the queue at the switch grow. >xpress your answer in 6.yy format. >g. %.'9 (to indicate %.'9G).
Thernet !witching ?ata #ateM % port bridge supports %1bps on each port and interconnects >thernet segments. The total number of hosts the bridge supports is 7% across all ports. % " points % " attempts •
;hat is the maximum data rate a backlogged host can achieve5 ssume that there is at least one host per port. >xpress your answer in 1bps. Format 6 (e.g. 9)
% " 7 points % " ' attempts •
;hat is the minimum data rate a backlogged host can achieve5 ssume that there is at least one host per port. >xpress your answer in kbps. Format 6 (e.g. '%)
/earning bridge % " points % " attempts •
;hat would a learning bridge do when it receives a packet with destination address as =M=M=M=M=M=5 send it on all ports send it on all ports except the one on which received drop it look up forwarding table to determine correct port
!ee Figure. The bridges (square boxes) are learning bridges with initial empty tables. The following transmissions take place. sends to I ? sends to & & sends to > sends to ? > sends to & % " ' points % " ' attempts •
fter the above transmissions, which bridges still do not learn where > is5 ll learn & &7 &,&7 % " B points % " ' attempts •
fter the above transmissions, which bridge has least amount of entries in its learning table5 !pecify the name of the bridge. >.g. &
!TE /ayer % " points % " attempts •
t what layer does the spanning tree protocol work5 ?ata+/ink etwork
Transport pplication
1ulti+paths % " points % " attempts •
:n some cases, the spanning tree protocol produces multiple paths between the same two / segments. Les o
!panning Tree. !ee Figure % " B points % " ' attempts :n the >xtended / topology after running the spanning tree protocol, which all ports will Kde$nitelyK or KpossiblyK be disabled5 1ark all that apply. &7+ &B+ &'+I &'+F &'+> &9+>
Hint KEossiblyK because there are two equally valid options for root port. % " 7 points % " ' attempts •
!uppose bridge & failed. ?etermine which ports will Kde$nitelyK or KpossiblyK be disabled after the spanning tree protocol kicks in again5 1ark all that apply. &'+>
&B+ &9+I &9+>
pplication /ayer ?! #ecords % " points % " attempts Ian hosts within the same subnet (sharing the same subnet id) be in di=erent ?! domains5 Les
No
% " points % " attempts :n ?! records, can many names map to the same value5 :f so, specify an example. :f not, why not5
$e o
>mail prottocols acation daemons % " points % " attempts •
acation daemons work by sending an automated out of oNce emails. ;ho implements these daemons, user agent or message transfer agent5 ote that one does use user+agent to set them up but who actually sends (not forwards) the reply emails5 Sser+agent
&e'ge-tr'ner-'gent ;eb mail protocol % " points % " attempts •
;hat protocol does web mail (at the user+end) use5 !1TE E0E7 :1E
HTT
DTTE 1essagesM user wishes to download a webpage which has 7 embed images within it. ssume that request message siAe is %%& and TIEKs 1!! is B%%&. lso assume that all obPects reside on the same server. % " ' points % " ' attempts •
Dow many DTTE request messages are sent5 Format 6 (e.g. 9)
% " points % " attempts •
Dow many DTTE response messages are sent5 Format 6 (e.g. ')
on+persistent connection % " points % " attempts •
Ian a non+persistent connection carry multiple DTTE request messages in the same TIE segment5 Les
No
TIE connections
% " points % " attempts •
!uppose a base DT1/ page with two obPects embed within have to be retrieved from the same server. Dow many TIE connections will be opened assuming DTTE operates over persistent connections5 Format 6 (e.g. 9)
#TTsM !uppose a user wants to access a webpage with 9 embed obPects all residing on the same server. ssume that length of the request messages is %%& and all obPects are under '%%&. ssume a TIE 1!! of B%%&. % " ' points % " ' attempts •
Dow many #TT would it take to display the webpage fully assuming non+persistent DTTE with no parallel TIE connections5 Format 6 (e.g. '8)
% " points % " attempts •
Dow many #TT would it take to display the webpage fully assuming non+persistent DTTE with parallel TIE connections5 Format 6 (e.g. '8)
% " points % " attempts •
Dow many #TT would it take to display the webpage fully assuming persistent DTTE with pipelined TIE connections5 Format 6 (e.g. '8)
;eb caching % " points % " attempts •
;hen requesting web obPects, does web caching reduce delay for all obPects or only for some of the obPects5
*!! o+,et
!ome obPects
