AIT 622 - Network Architecture and Protocols Fall 2011

divisionimpossibleNetworking and Communications

Oct 24, 2013 (3 years and 10 months ago)

367 views

Name:______
____________________
_________


AIT 622

-

Network Architecture and Protocols

Fall

2011


Exam 2



MULTIPLE CHOICE

(
2

points each)


Circle the letter of the
one

best answer for each question.



1
. In TCP, the timer value used by a sender to dete
rmine when to retransmit a segment:


a.

is potentially different for each TCP connection

b.

is revised as the performance of a connection changes

c.

is made up of three components: the RTT estimate, the estimate of the variance in delay,
and the estimate of the clo
ck differences between the two machines

d.

all of the above

e.

a and b above



2
.
With regard to
ICMP:


a.

it is an error
correcting

mechanism (versus only error
reporting
)

b.

it allows machines to send error or control messages to other machines

c.

machines can only se
nd ICMP messages to machines on a directly connected network

d.

all of the above

e.

b and c above



3
. UDP:


a.

violates the layering principle by constructing the pseudo
-
header

b.

uses IP to transport messages between machines

c.

provides protocol ports that are used t
o distinguish among multiple programs executing on
a single machine

d.

all of the above

e.

a and b above

f.

b and c above



4
. The ICMP
Echo Request and Reply

(ping) message:


a.

is used to assist in flow control

b.

informs hosts that there is a better route available

c.

i
s used to test whether a destination is reachable and responding

d.

assists in clock synchronization

5
. An exterior gateway protocol:


a.

is used to pass information among routers
inside

an autonomous system

b.

is used to pass information among routers
across

two
autonomous systems

c.

is implemented by BGP

d.

all of the above

e.

a and c above

f.

b and c above



6
.
Proxy ARP
:


a.

can be used in networks even if ARP is not used for address resolution

b.

allows a single network prefix to be used for two physical networks

c.

is no longer
used because of classless addressing

d.

all of the above



7
. The 32
-
bit subnet mask: 11111111 11111111 11111111 00000000



a.
is only valid

for a Class A network


b.
is only valid

for a Class B network


c.
is only valid

for a Class C network


d.
could be

valid for Class A, B,
or

C networks



8
. Many texts differentiate between flow control and congestion as follows:
flow control

involves
preventing senders from overrunning the capacity of receivers.
Congestion control

involves
preventing too mu
ch data from being injected into the network, thereby causing routers or links to
become overloaded. What scheme does TCP use for
flow control
?


a. the sliding window protocol with a variable window size

b. positive acknowledgements

c. positive acknow
ledgements with retransmission

d. TCP does not provide flow control



9
. TCP is able to provide reliable delivery service even though the underlying communication
system is unreliable packet delivery because:


a. packets never really get lost

b. it use
s positive acknowledgements with retransmission

c. congestion control algorithms eliminate unreliability

d. the application layer takes care of the problem

e. all of the above





10
. Subnetting:



a. allows a single network address to span multiple p
hysical networks


b. was made obsolete by the introduction of supernetting


c. is only useful for Class B networks


d. all of the above



11
. Data structures and search algorithms used with
classful

addresses do not work with
classless

addresses becaus
e:



a. classless addresses do away with the need for speed in finding the next hop information

b. classless addresses are not self
-
identifying

c. classless addresses use less bits

d. classless addresses use more bits

e. the statement is not true; th
e same data structures and search algorithms can be used with


both classeful and classless addressing



12
. The “extra hop problem” is caused by:



a. congestion at one or more routers along a path


b. limiting the number of routers that can partic
ipate in a routing protocol and



having routers outside the group use a member of the group as a default route


c. silly window syndrome


d. hidden networks




13
. Link
-
state routing:


a.
does
not

scale

as well as the D
-
V algorithms

b.
requires e
ach participating router to have complete topology information

c.
is

guaranteed to converge

since the computation is done locally

d. suffers from the large link status messages that must be propagated

e. all of the above

f. b and c above



14
. Silly
window syndrome is:



a. caused by Karn’s algorithm


b. prevented on the send side by Karn’s algorithm


c
. characterized by each ack advertising a small window size and each segment



carrying a small amount of data


d
. a situation in which the adv
ertised window size grows without bound (that is, very large)


e
.
b and c above




15. With regard to Distance
-
Vector (DV) routing algorithms:


a.

their main disadvantage is that they do not scale well

b.

routers periodically exchange
only the portion of their

routing table that has changed

since
the last exchange

c.

they have been made obsolete by the link state algorithms

d.

all of the above





MATCHING
:

(1
0

points
; 1.25 points each
) For each statement below, put the letter of the correct
term in the blank before

the statement. A term will be used only once; not all terms will be used.



_____
1
.
Group of networks and routers controlled by a single administrative authority


_____
2
.
A condition of severe delay caused by an overload of datagrams at one or more rou
ters


_____
3
.
One of the first uses of classless addressing


_____
4
.
Used to pass network reachability information
between

two autonomous systems


_____
5
.
Alternative to the
tail
-
drop

scheme for managing queue overflow at routers


_____
6
.
Policy used b
y routers for discarding datagrams that simply says if the input queue is filled
when a datagram arrives, discard the datagram


_____
7
.
Technique used to shrink the congestion window during times of congestion


__
___8
.
Technique used to increase the size

of the congestion window after congestion


has cleared out




Terms:


A.
Autonomous System


B.
LAN






C.
Tail
-
drop


D.
ICMP




E.
ARPANET




F.

Sliding Window


G.
Proxy ARP



H.
Split Horizon




I.
C
ongestion


J.
Hold
-
down





K.
Multiplicative Decrease


L.
Poison Reverse


M.

Retransmissions



N.
RTT





O.
RED



P.
Subnetting




Q.

Transparent Routers



R.

EGP



S.
Slow
-
start





T.
Supernetting




U.
R
IP

Short Answer:

(
10

points each) Answer
any 6

of the following 7 questions. If more than 6 are
answered, the first 6 will be graded!! (unless you specifically indicate which questions to grade)

READ each question carefully and answer completely and co
ncisely!!



1. Briefly describe how the Random Early Detection (RED) policy works (that is, how it decides
when to discard a datagram).




















2.
Discuss the relationship between classless addressing, CIDR, and supernetting.













3
. W
hat is the purpose of the Nagle algorithm? Describe how it operates.





4
. Consider an existing distance
-
vector table for router
K

and a just
-
received update from router
J
,
both as shown below.
For each change

that should occur in the existing table f
or router
K
, give:


-

the change to be made (add route to
X
, update route to
X
, etc.),


-

the reason for the change, and


-

what the resulting entry will be.

Assume that router
J

is only one hop from router
K
.


NOTE: The change triggered by the inform
ation for Net 3 is shown as an example. Fill in all
other
changes

that would be made based on the incoming update message from router
J
.


YOU ONLY NEED TO SHOW THE
CHANGES

TO ROUTER K’s TABLE!



Destination

Distance

Route


Destination

Distance


Net
4

0

direct



Net 3

7


Net 6

0

direct



Net 6

1


Net
9

7

Router
D



Net
9

6


Net 10

5

Router S



Net 10

3


Net 1
2

4

Router J



Net 14

8


Net 1
5

6

Router
J



Net 15

6


Net 16

6

Router J



Net 16

5

Ex
isting table for router
K


Incoming update message
from router
J


Change to be
made

Reason for the change

Resulting Entry

Destination

Distance

Route


Add route to 3











J has a route; K does not









Net 3








8








Router J








EXAMPLE
:

5
. Suppose you are the network administrator for a network with the address 220.98.32.0. Further
suppose you want to use subnetting to allow for at least 10 different physical networks with a
minimum of 12 PC’s on each. Determine the subnet mask that wou
ld allow this and state how many
subnets and hosts per subnet would actually be possible with the chosen mask.



a. Proposed subnet mask (dotted quad or binary):









b. Maximum subnets allowed based on the mask:





c. Maximum hosts per subnet all
owed based on the mask:





6. What is the purpose of Karn’s algorithm? Describe how it operates.















7
.

Briefly describe the three
-
way
-
handshake process used to open a TCP connection
.









EXTRA

CREDIT:


(
5

points)
Describe why UDP us
es the pseudo
-
header to verify that the UDP datagram has reached
the correct destination
,

even though the IP layer has already used the destination IP address to
ensure

that the datagram is addressed to the receiving machine.


















Potential Ref
erences

F
F
i
i
g
g
u
u
r
r
e
e


1
1
2
2
.
.
1
1
5
5




























































F
F
i
i
g
g
u
u
r
r
e
e


9
9
.
.
5
5