Download Communication Engineering Networks - 2000 2001 Exam - Electrical Engineering and more Exams Electrical Engineering in PDF only on Docsity!
S343 06/09 /
TH E MANCH ESTER M ETR O PO LITAN UNIVER SITY
FACULTY O F SCIENCE AND ENGINEER ING
D EPA R TMENT O F ENGINEER ING AND TECH NO LO GY
SESSIO N 2000/
Exam ination for th e BEng (H O NS) ELECTR ICA LAND ELECTR O NIC ENGINEER ING BEng (H O NS) CO MMUNICA TIO N AND ELECTR O NIC ENGINEER ING BEng (H O NS) CO MPUTER A ND ELECTR O NIC ENGINEER ING BEng (H O NS) ELECTR O NIC ENGINEER ING W ITH M A NA GEMENT BEng (H O NS) ELECTR O NIC ENGINEER ING (PA R T-TIME) FINA LYEA R
UNIT 64EE3018: CO M M U NICATIO N ENGINEER ING AND NETW O R K S
Tuesday 15 May 2001
9 .30 am to 12.30 pm
Instructions to Candidates
A nsw er FIVE questions.
A llquestions carry equalm ark s.
Th e follow ing supplem entary answ er sh eets are provided separately and sh ould b e h anded in w ith your script: Figure Q 4b , Figure Q 4c and Figure Q 8.
06/09 /02 continued
- (a) Briefly explain w ith th e aid of sk etch es, w h y pulses h aving a raised cosine spectrum are frequently em ployed in pulse transm ission system s. [4]
(b) O n-off NR Z pulses of am plitude A 1 =20μV and A 1 =0V respectively are received togeth er w ith w h ite Gaussian noise at th e output of a ch anneland are passed into a decision gate. Th e th resh old for detection is set m idw ay b e tw een th e logic 1 and logic 0 levels. Th e transm ission b andw idth of th e ch annelextends from 0 to 100MH z, w ith infinite attenuation of frequencies ab ove 100MH z. Th e noise at th e output of th e ch annelis assum ed to b e of uniform spectraldensity, of double-sided intensity 10-20V^2 /H z, over th e transm ission b andw idth.
Explaining in detailallof th e steps tak en, evaluate th e error prob ab ility and th e average tim e b e tw een errors (assum ing th at th e pulses h ave a raised cosine spectrum ) w h en transm itting at th e m axim um signalling speed. [16]
It m ay be assum ed th at th e prob ab ility density function of th e noise is given by:
p v
e
v
( ) =
− 2 2 2 2
σ σ π and th at th e ch annelis b inary sym m etric.
06/09 /02 continued
- Measurem ents from m onitors in th e w ide-area netw ork sh ow n in Figure Q 3 are given in Tab le Q 3, togeth er w ith th e pack et routing inform ation. Th e m easurem ents are in th e form of average pack et rates given in pack ets/s and th e average pack et length is 10 k bytes. It m ay be assum ed th at allarrivals are Poisson and th at pack et transm ission tim es on alllink s are exponentially distributed.
D e stination A B C D E F A 0 10 A B
A C
A BD
A CE
A CF B 10 BA
BA C
BD
BD F E
BD E
Source C 12 CA
CA B
CD
CE
CF D 8 D BA
D B
D C
D FE
D F E 14 ECA
EFD B
EC
EFD
EF F 10 FCA
FD B
FC
FD
FE
(a) Find th e average delay on each of th e netw ork link s and th e num b e r of pack ets in th e system at th e B to D link. [8]
(b ) Sh ow th at th e m axim um -flow is equalto th e m inim um -cut for th e netw ork sh ow n in Figure Q 3. [12]
A
Manch ester
R eading
Birm ingh am Cam b ridge
Leeds
London
C D
E
B
F
8Mb it/s
8Mb it/s 8Mb it/s
8Mb it/s 8Mb it/s
8Mb it/s
155Mb it/s
155Mb it/s
Netw ork Structure
Figure Q 3
Traffic Matrix
Tab le Q
06/09 /02 continued
- (a) Explain, w ith reference to a h istoricalperspective, w h y cell-b ased netw ork s are used as m ultim edia netw ork s. [5]
(b ) D e scrib e h ow th e organisation of celltransfer tak es place b e tw een a source station and a destination station in an asynch ronous transfer m ode localarea netw ork. [5]
(c) A segm ent of an asynch ronous transfer m ode netw ork is sh ow n in Figure Q 4a. Th e num b e rs adjacent to each R CU/sw itch are th e port identifier. Sem iperm anent virtualcircuits are to b e set up by netw ork m anagem ent b e tw een m ultim edia w ork stations A ,B,C and D to th e SCP for on-dem and signalling and inform ation transfer.
Th e follow ing assum ptions m ay be m ade:
(i) VP-only sw itch ing is used w ith in th e netw ork sw itch es and R CU's, (ii) th at th e VPI/VCI field is used to identify specific calls/stations, (iii) virtualpath s sh ould b e filled starting from th e low est upw ards.
D e rive typicalrouting tab le entries using Figure Q 4b and Q4c for R CU 1, R CU 2, SW 1 and SW 2, SCP and CLS to provide th ese connections. [10]
A
B
D
C
R CU 1
R CU 2
SW 1 SW 2^ SCP
CLS Server
ATM Netw ork Segm ent
Figure Q 4a
SW 2 In O ut In O ut Port VPI VCI Port VPI VCI Port VPI VCI Port VPI VCI SC CC CC CC CLS SC CC CC CC CLS SC CC CC CC CLS SC CC CC CC CLS CLS' CLS'
CLS In O ut In O ut VPI VCI VPI VCI VPI VCI VPI VCI
SCP
Signalling Callch annels ch annel (^) In O ut VPI VCI Port VPI VCI Port VPI VCI
R egistration Num b e r:________________________
Figure Q 4c
06/09 /02 continued
- (a) A sch em atic diagram of a datacom m unication netw ork is sh ow n b e low in figure Q 5. Explain h ow a data m essage w ould b e sent using IP pack ets from nod e A , over an Internet w ith interm ediate routers, to a rem ote node B. Your answ er sh ould consider in som e detail, and w ith th e aid of a diagram , h ow fragm entation and addressing at th e different layers is used as th e data passes th rough th e netw ork from source to destination nodes. [12]
Data link Physical
Network Data link Physical
Network Data link Physical
Data link Physical
Network
Transport Transport
Session Session
Presentation Presentation
Application Application
connection between end users
Figure Q 5
A continuous series of 1000-b it fram es is to b e transm itted w ith a negligib le b it error rate over a datacom m unication link using an idle R Q e rror controlprotocol. You m ay assum e th at th e A ck now ledgem ent fram e transm ission tim e and th e processing tim e at each node are negligib le com pared to th e Inform ation or I -fram e transm ission tim e (tfram e) and th e propagation tim e (tprop) across th e datalink. Tak e th e velocity of propagation of free-space link s to b e 3 x 10^8 m /s and cab le link s to b e 2 x 10^8 m /s.,
(b ) D e term ine th e link efficiency for th e follow ing types of data link assum ing a data transm ission rate of 1 Mb it/s.
(i) A tw isted pair cab le 1 k m in length. (ii) A satellite link of 50000k m. [6]
(c) State w h at generalconclusions can b e d raw n from th e ab ove calculations. [2]
Node A (^) Node B
06/09 /02 continued
- Th e follow ing opticalfib re com m unication system is required.
A 100 K m opticalfib re link to transm it a 1Gb it/s NR Z data signal. Th e fib re link is to com prise 10k m separate length s of single m ode fib re, spliced togeth er w ith fusion splices. A ssum e th at th e fusion splices h ave an average loss of 0.02dB for single m ode fib re. A n overallsystem safety m argin of 3dB is required.
Th e follow ing Tab les provide th e relevant data inform ation on th e com ponents th at are availab le to th e designer of th is system.
Transm itter type
O perating λ (nm )
SpectralLinew idth σtx (nm )
Launch pow er (dBm )
R ise tim e (ns) D FB laser 1300 0. 3 5 0. 07 D FB laser 1550 0. 3 5 0. 05
9 /125 singlem ode fib re
O perating λ(nm )
A ttenuation (dB.k m -1)
Intram odal dispersion D int (ps.nm -1.k m -1) 1300 0. 8 3. 5 1550 0. 3 10
R eceiver type O perating λ (nm )
Sensitivity (dBm )
R ise tim e (ns) A PD 1300 to 1550nm -55 0. 5
(a) D e term ine th rough suitab le calculation w h eth er th e fib re system w ill operate correctly at an operationalw avelength of 1300nm. State any assum ptions m ade in your analysis. [12]
(b ) Given th at th e ph ysicalfib re link w ith in th e overallsystem is fixed and cannot b e replaced, explain w ith justification th e m ost appropriate ch anges th at could b e m ade to th e design to im prove th e overallsystem. [6]
(c) State b riefly w h at th e im plication of th e ab ove design h as in generalterm s regarding th e operation of long h aulopticalfib re telecom m unication link s. [2]
S343 06/09 /
- (a) Explain w h y sim ple analog space sw itch ing is no longer considered a valid tech nique in m odern teleph one exch anges. [4]
Th e tab le b e low sh ow s th e connections required betw een th e input lines A , B, C and output lines P, Q and R (for th e T-S-T netw ork sh ow n b e low over a period com prising 4 tim e slots).
Input line A 1 A 2 A 3 A 4 B1 B2 B3 B4 C1 C2 C3 C O utput line R 4 Q 4 P 2 P 3 P 4 P 1 R 1 R 3 Q 3 R 2 Q 1 Q 2 Fram ing D elay ( in tim e slots)
It is k now n th at th e stage 1 input tim e sw itch is set to provide a fixed tim e delay of 1 tim e slot for A LLinput lines.
(b ) Com plete and attach to your answ er book th e tim e sw itch connection tab les in th e figure Q 8 to give a m inim um overalldelay th rough th e sw itch netw ork. A lso clearly m ark on figure Q 8 th e space sw itch cross connects for tim e slot 2. [8]
(c) D e scrib e h ow em erging A sym m etric D igitalSubscrib e r Line (A D SL) tech nology operates, sh ow ing w ith th e aid of a suitab le diagram h ow it can b e com b ined in th e localloop w ith th e existing Plain O ld Teleph one Service (PO TS). [8]
END
