ECE374:

 First  Midterm                                                                                                                                                                                                                                        1  
 
Second  Midterm  for  ECE374  
 04/08/15  
Solution!!  
 
Instructions:  
a. Put  your  name  and  student  number  on  each  sheet  of  paper!  
b. The  exam  is  closed  book.  
c. You  have  90  minutes  to  complete  the  exam.  Be  a  smart  exam  taker  -­‐  if  you  
get  stuck  on  one  problem  go  on  to  another  problem.  Also,  don't  waste  your  
time  giving  irrelevant  (or  not  requested)  details.  
d. The  total  number  of  points  for  each  question  is  given  in  parenthesis.  There  
are  100  points  total.  An  approximate  amount  of  time  that  would  be  
reasonable  to  spend  on  each  question  is  also  given;  if  you  follow  the  
suggested  time  guidelines,  you  should  finish  with  10  minutes  to  spare.  The  
exam  is  90  minutes  long.  
e. Show  all  your  work.  Partial  credit  is  possible  for  an  answer,  but  only  if  you  
show  the  intermediate  steps  in  obtaining  the  answer.  If  you  make  a  mistake,  
it  will  also  help  the  grader  show  you  where  you  made  a  mistake.  
f. Good  luck.  
 
Problem   Max.  Points     Points  
1   32    
2   20    
3   24    
4   24    
Total   100    
ECE374:  First  Midterm                                                                                                                                                                                                                                  2  

Problem  1:  (Quickies  32  Points  (4  each),  25  minutes)  

a. Briefly explain at what point the TCP congestion control algorithm switches from
the slow-start to the congestion avoidance phase? Is it possible that the algorithm
is also switching back from the congestion avoidance to slow-start. Briefly
explain your answer.
Answer: This transition happens when the slow-start threshold (sshtresh) is
reached. Yes, in the case of packet loss or time out.

b. Specify the IP network address, the host range, and the subnet mask for a subnet
that can contain up to 30 clients. What is the maximum number of hosts per
subnet if you have 24 subnet mask bits?  
Answer:  192.168.0.0;  Host  range:  192.168.0.1  -­‐  192.168.0.30;  Mask  bits:  27;  
254;  

c. Suppose  a  web  server  runs  in  Host  C  on  port  80.  Suppose  this  Web  server  
uses  persistent  connections,  and  is  currently  receiving  requests  from  two  
different  Hosts,  A  and  B.  Are  all  of  the  requests  being  sent  through  the  same  
socket  at  Host  C?  If  they  are  being  passed  through  different  sockets,  do  both  
of  the  sockets  have  port  80?  Explain  briefly.  
Answer:  For  each  persistent  connection,  the  Web  server  creates  a  separate  
“connection  socket”.  Each  connection  socket  is  identified  with  a  four-­‐tuple:  
(source  IP  address,  source  port  number,  destination  IP  address,  destination  port  
number).  When  host  C  receives  and  IP  datagram,  it  examines  these  four  fields  in  
the  datagram/segment  to  determine  to  which  socket  it  should  pass  the  payload  
of  the  TCP  segment.  Thus,  the  requests  from  A  and  B  pass  through  different  
sockets.  The  identifier  for  both  of  these  sockets  has  80  for  the  destination  port;  
however,  the  identifiers  for  these  sockets  have  different  values  for  source  IP  
addresses.  Unlike  UDP,  when  the  transport  layer  passes  a  TCP  segment’s  
payload  to  the  application  process,  it  does  not  specify  the  source  IP  address,  as  
this  is  implicitly  specified  by  the  socket  identifier.  
 
d. Do  routers  have  IP  addresses?  If  so,  how  many  IP  addresses  and  does  each  IP  
address  have  to  belong  to  the  same  or  a  different  subnet?    
Answer:  Yes;  As  many  as  the  router  has  interfaces;  Each  has  to  belong  to  a  
different  subnet;    
 
e. Briefly  explain  how  IPv6  datagrams  can  be  tunneled  over  an  IPv4  backbone.  
You  can  use  a  simple  network  (as  sketched  below)  that  consist  of  4  routers  
that  are  daisy  chained  to  explain  how  the  tunneling  works.  
IPv6 ---- IPv4 ---- IPv4 ---- IPv6  
Answer:  IPv6  -­‐-­‐-­‐-­‐  IPv6  encapsulated  in  IPv4  -­‐-­‐-­‐-­‐  IPv6  encapsulated  in  IPv4  -­‐-­‐-­‐-­‐  
IPv6  
 
f. Why  is  an  ARP  query  sent  within  a  broadcast  frame?  Why  is  an  ARP  response  
sent  within  a  frame  with  a  specific  destination  MAC  address?  
Answer:  Sender  does  not  know  MAC  address  of  host  that  has  specific  IP  address,  
 ECE374:  First  Midterm                                                                                                                                                                                                                                        3  
 
yet.  Learns  MAC  of  host  from  incoming  ARP  request,  no  broadcast  necessary.  
 
g. Briefly  describe  how  Ethernet’s  exponential  backoff  works.  What  is  one  
reason  why  Ethernet’s  exponential  backoff  might  be  better  than  randomizing  
retransmission  attempts  over  a  fixed-­‐length  time  interval?    
Answer:  Ethernet  maintains  an  interval  of  time  T  over  which  is  will  randomize  
when  it  will  attempt  a  retransmission.  After  each  collision  for  the  same  packet,  
it  doubles  the  length  of  T  up  to  some  fixed  max.  This  is  better  than  just  a  single,  
fixed  value  of  T  since  when  there  are  a  lot  of  collisions  the  interval  over  which  
randomization  is  done  will  be  large,  allowing  just  one  node  to  successfully  
being  transmitting.  When  there  are  only  a  small  number  of  colliding  nodes,  the  
retransmission  will  be  randomized  initially  over  a  small  T,  allowing  a  node  to  
transmit  more  quickly.  

h. Briefly describe the difference between ad-hoc and infrastructure mode in

wireless networks. Give one example advantage of ad-hoc over infrastructure
mode.
Answer: Ad-hoc: node-to-node communication; Infrastructure: Nodes do only
communicate with base station and not with other nodes. Fault tolerance; if base
station fails, no more communication is possible.
 
 
Problem  2:  Congestion  Control  &  Avoidance  &  Routing  (20  Points,  20  minutes)  
 
a. (2  Points)  Why  do  you  think  TCP  does  not  include  RTT  measurements  for  
retransmitted  packets  in  the  SampleRTT?  
Answer: Let’s  look  at  what  could  wrong  if  TCP  measures  SampleRTT  for  a  
retransmitted  segment.  Suppose  the  source  sends  packet  P1,  the  timer  for  P1  
expires,  and  the  source  then  sends  P2,  a  new  copy  of  the  same  packet.  Further  
suppose  the  source  measures  SampleRTT  for  P2  (the  retransmitted  packet).  
Finally  suppose  that  shortly  after  transmitting  P2  an  acknowledgment  for  P1  
arrives.  The  source  will  mistakenly  take  this  acknowledgment  as  an  
acknowledgment  for  P2  and  calculate  an  incorrect  value  of  SampleRTT.  
 
b. (2 Points) Does UDP require a mechanism to estimate the RTT between sender
and receiver? Briefly explain your answer?
Answer: No, UDP does not perform retransmissions and, thus, does not require
an RTT-based timeout timer.  
 
Consider  sending  a  large  file  from  a  host  to  another  over  a  TCP  connection  that  has  
NO  loss  
c. (5  Points)  Suppose  TCP  uses  AIMD  for  its  congestion  control  without  slow  
start.  Assuming  cwnd  increases  by  1  MSS  every  time  a  batch  of  ACKs  is  
received  and  assuming  approximately  constant  round-­‐trip  times,  how  long  
does  it  take  for  cwnd  to  increase  from  5  MSS  to  10  MSS?  
Answer: It takes 1 RTT to increase CongWin to 6 MSS; 2 RTTs to increase to 7
ECE374:  First  Midterm                                                                                                                                                                                                                                  4  

MSS; 3 RTTs to increase to 8 MSS; 4 RTTs to increase to 9 MSS; 5 RTTs to

increase to 10 MSS.  
d. (5 Points) What is the average throughput (in terms of MSS and RTT) for this
connection up through time 6 RTT?
Answer: In the first RTT 5 MSS was sent; in the second RTT 6 MSS was sent; in
the third RTT 7 MSS was sent; in the fourth RTT 8 MSS was sent; in the fifth RTT,
9 MSS was sent; and in the sixth RTT, 10 MSS was sent. Thus, up to time 6 RTT,
5+6+7+8+9+10 = 45 MSS were sent (and acknowledged). Thus, we can say that
the average throughput up to time 6 RTT was (45 MSS)/(6 RTT) = 7.5 MSS/RTT.  
 
 
 

 
Figure  1  

Consider  Figure  1  for  the  following  problem.  

e. (2  Points)  In  this  case,  the  two  networks  A  and  B  are  connected  through  
network  C  to  the  rest  of  the  Internet.  What  network  address  or  addresses  for  
A  and  B  are  advertised  by  the  gateway  router  (that  connects  C  to  the  
Internet)?  Specify  network  address  and  subnet  mask?  
Answer: 128.119.0/16  
f. (2 Points) Assume that the router that connects C to both A and B has interface 1
connected to A and interface 2 connected to B. For these two networks, specify
the forwarding table entries in the router of network C.
Answer:128.119.0/17 | Interface 1, 128.119.128/17, Interface 2.  
g. (2 Points) Now assume that a host somewhere in the Internet (the host is not
located in networks A, B, and C) sends a packet to i) a host in network A and ii)
to a host in network B. Explain how the router (from problem b.) in network C
determines which interface to send each of these packets.
Answer: longest prefix matching.  
 
 
Problem  3:  DHCP  &  NAT  &  Routing  (24  Points,  20  minutes)  
This  problem  is  concerned  with  DHCP  and  NAT.  Let’s  first  focus  on  DHCP.    
a. (4  Points)  What  four  pieces  of  information  does  a  DHCP  client  obtain  from  
the  DHCP  server  after  a  successful  communication  between  the  two?    
Answer: IP address, IP address of first hop router, name and IP address of DNS
server, network mask.  
 ECE374:  First  Midterm                                                                                                                                                                                                                                        5  
 
b. (4  Points)  What  are  the  destination  IP  and  MAC  address  in  the  discover  
request  sent  by  the  DHCP  client?  What  is  the  source  IP  address  and  MAC  
address  of  the  offer  message  sent  by  the  DHCP  server?  
Answer: IP = 255.255.255.255 MAC: FF:FF:FF:FF:FF:FF, IP = IP address of
DHCP server, MAC = MAC address of DHCP server.  
 
Now  let’s  look  at  how  NAT  work.    
 

 
Figure  2  

   
 
 
 
c. (2  Points)  Specify  IP  addresses  for  all  interfaces  of  the  local  network.  Also  
assign  MAC  addresses  to  these  interfaces.  
Answer:

 
d. (6  Points)  Now  assume  that  the  top-­‐most  client  in  the  local  network  sends  an  
HTTP  request  to  http://www.umass.edu  (IP  addr:  128.119.103.148)  and  the  
local  port  for  the  TCP  connection  for  this  HTTP  request  is  7000.  Fill  out  the  
NAT  table  below  with  the  information  that  would  be  entered  in  that  table  
after  the  HTTP  request  has  been  forwarded  from  the  NAT  router  into  the  
WAN.  
In  addition,  specify  destination  and  source  IP  addresses  and  ports  for  
ECE374:  First  Midterm                                                                                                                                                                                                                                  6  

segment  1,  2,  3,  and  4  as  indicated  in  Figure  2.  
 
NAT  translation  table  
WAN  side  address/port   LAN  side  address/port  
128.119.103.148,  5000   10.2.0.2,  7000  
128.119.103.148,  5001   10.2.0.3,  7000  
128.119.103.148,  5002   10.2.0.4,  7000  
 
Answer:  
e. (2  Points)  Also  specify  MAC  source  and  destination  address  for  1  and  2.  
Answer: BB:BB:BB:BB:BB:BB, AA:AA:AA:AA:AA:AA  
 
When  the  packet  leaves  the  NAT  router  it  needs  to  be  routed  through  the  Internet  to  
its  destination  (the  UMass  web  server).  For  this  problem,  let’s  assume  the  network  
shown  in  Figure  3.    
 

 
Figure  3  

f. (6  Points)  Assume  node  u  represents  the  NAT  router  and  node  z  is  the  router  
the  UMass  web  server  is  connected  to.  By  applying  Dijkstra’s  shortest  path  
algorithm  fill  out  the  table  below  and  determine  the  shortest  path  from  u  to  z.  
Step   N’   D(v),  p(v)   D(w),  p(w)   D(x),  p(x)   D(y),  p(y)   D(z),  p(z)  
0   u            
1              
2              
3              
4              
5              
 
   
 
 
 
 
 
 ECE374:  First  Midterm                                                                                                                                                                                                                                        7  
 
Answer:  
 
Step   N’   D(v),  p(v)   D(w),  p(w)   D(x),  p(x)   D(y),  p(y)   D(z),  p(z)  
0   u   2,  u   ∞   1,  u   ∞   ∞  
1   ux   2,u   4,x     2,x   ∞  
2   uxy   2,u   3,y       4,y  
3   uxyv     3,y       4,y  
4   uxyvw           4,y  
5   uxyvwz            
 
     
ECE374:  First  Midterm                                                                                                                                                                                                                                  8  

 
Problem  4:  Ethernet  (24  Points,  25  minutes)  
The  first  part  of  this  problem  focuses  on  802.3.  
a. (5  Points)  In  the  scenario  shown  in  Figure  4,  A  is  sending  a  frame  to  A’  and  
after  receiving  that  frame,  A’  sends  a  frame  back  to  A.  Explain  the  steps  that  
are  carried  out  by  A,  A’,  and  the  switch  to  populate  its  table  such  that  packets  
from  A  can  be  switched  to  A’  and  vice  versa.  
 
 

Figure  4  

Answer:    
MAC  Addr   Interface   TTL  
A   1   60  
A’   4   60  
Step  1:  A  sends  frame  to  switch  on  port  1,  switch  know  that  it  can  reach  A  on  
that  port.  
Step  2:  Switch  floods  packet  out  all  ports  except  1  because  it  does  not  no  to  
which  port  A’  is  connected  to.  
Step3:  A’  replies  and  switch  learns  that  A’  is  connected  to  port  4.  
 
b. (3  Point)  Does  each  port  on  the  switch  have  to  have  its  own,  unique  MAC  
address?  Briefly  explain  your  answer.  
Answer:  No.  Since  each  node  is  directly  connected  to  the  switch  MAC  addresses  
are  not  required  at  the  switch  ports  (a  packet  sent  from  a  host  will  always  end  
up  at  the  switch  port  that  connects  to  the  host).  Actually,  the  switch  is  
transparent  for  the  hosts.  E.g.,  in  the  case  of  A  sending  a  frame  to  A’,  the  frame  
only  carries  A  and  A’  MAC  addresses  and  no  information  about  the  switch  or  its  
ports.  
 
c. (2  Points)  In  Figure  5,  A  is  sending  an  ARP  request  to  determine  the  MAC  
address  it  should  use  for  a  frame  destined  to  B.  What  MAC  address  will  be  
returned  to  A  in  the  ARP  reply.  Briefly  explain  your  answer.  
 ECE374:  First  Midterm                                                                                                                                                                                                                                        9  
 
Answer:20:FF:3A:BC:01:4E.  Since  B  is  not  in  the  same  switching  domain  as  B  
frames  have  to  be  sent  to  the  left  interface  of  R1.    
d. (2  Points)  How  many  interfaces  of  all  the  ones  shown  in  Figure  5  receive  the  
original  ARP  request  sent  by  A  and  what  is  the  destination  MAC  address  on  
the  frame  containing  the  ARP  request?  
Answer:  6,  FF:FF:FF:FF:FF:FF  
 
C

BB:89:34:E7:01:3B

1
(3)
Switch 4 Switch 3
2
10:D4:E1:A8:97:F0 3 B
AA:10:F3:5C:01:04
R1
(1) (2)
Switch 1 1 3 Switch 2

A 20:FF:3A:BC:01:4E 2

 
Figure  5  

e.  (4  Points)  Let’s  assume  that  router  R1  shown  in  Figure  5  connects  two  IP  
subnets.  In  the  scenario  shown  in  this  figure  a  host  that  is  connected  to  either  
switch  1  or  4  cannot  change  from  its  current  subnet  to  the  other  subnet  
without  changing  its  physical  connection  to  either  switch  2  or  3.  Can  you  
think  of  a  solution  that  would  allow  this  host  to  stay  connected  to  its  original  
switch.  (Hint:  Think  VLAN).  If  so,  draw  a  sketch  or  your  proposed  solution.  
Answer:  

 
 
Let’s  now  focus  on  802.11    
ECE374:  First  Midterm                                                                                                                                                                                                                                  10  

f. (2  Points)  Briefly  explain  the  two  methods  through  which  a  wireless  802.11  
node  learns  about  access  points  it  can  communicate  with.  
Answer:  
i)  Passive  scanning,  beacon  frame  from  AP,  association  from  node,  reply  from  
AP.  
ii)  Active  scanning,  probe  request  from  node,  probe  responses  from  APs,  
association  request  from  node,  reply  from  AP.  
g. (2  Points)  Explain  how  the  ACK  transmitted  by  the  base  station  after  the  data  
transmission  from  the  host  signals  that  the  transmission  from  the  host  has  
finished  also  to  a  potential  hidden  terminal.  
Answer:  A  host  that  cannot  observe  the  transmission  of  another  host  but  is  in  
transmission  range  of  the  base  station  will  be  able  to  detect  the  ACK  and  no  
that  the  transmission  of  the  other  host  has  finished.  
h. (4  Points)  Figure  6  shows  a  typical  setup  for  a  infrastructure  mode  wireless  
network  where  the  BSS  includes  several  hosts  and  a  base  station.  The  later  is  
connected  to  the  Internet  via  a  route.  For  this  scenario  show  all  MAC  address  
in  correct  order  in  the  802.11  frame  that  is  transmitted  from  the  host  H1  to  
the  base  station  and  the  frame  transmitted  from  the  base  station  to  the  
router.  If  the  addresses  in  the  two  frames  are  different,  briefly  explain  why.  

  Figure  6  
 
 
Answer: