Professional Documents
Culture Documents
A
bit
has
the
unit
of
b
and
a
byte
has
the
what’s
being
carried
through,
depends
on
Multiply
M(x)
by
xk
where
C(x)
has
degree
k.
Let
unit
of
B,
a
byte
is
8
bits.
distance;
T
=
time
to
transfer
data,
the
result
be
T(x).
6
2. M
in
bandwidth
means
10 .
M
in
data
size
depended
on
bandwidth
and
total
data
size;
Find
remainder
R(x)
of
T(x)
/
C(x);
P(x)
=
T(x)
–
20
means
2 .
Q
=
queuing
delay
or
time
at
switches
and
R(x)
Network
requirements:
routers
TCP
makes
sure
each
message
is
received
(ack
Bandwidth:
Theoretical
Max,
#
of
bits/seconds
Example
of
finding
P(x):
M
=
10101101;
M(x)
=
x7
sent
back
and
forth)
which
gives
us
reliability
and
RTT:
round
trip
time,
2
x
P
+
x5
+
x3
+
x2
+
x0;
C(x)
=
x3
+
x2
+
1;
Since
the
manages
congestion
control.
Delay
x
bandwidth
=
(RTT/2)x
bandwidth
degree
of
C(x)
=
3,
we
multiply
M(x)
by
x3.
;
T(x)
A
network
is
2
or
more
linked
devices
capable
of
Throughput:
actual
data
carried
over
time,
#
of
=
x10
+
x8
+
x7
+
x5
+
x3
Now,
find
remainder
communication.
Facebook,
Comcast,
Internet,
bits/seconds
R(x)
when
dividing
T(x)
by
C(x)
(division
is
done
cell
networks,
radio,
and
your
brain
are
all
Problems:
M
=
Mbps
(throughput
or
bandwidth)
• Framing
Protocols
–
group
bits
into
with
subtraction
as
XOR):
examples
of
networks.
Important
network
10^6
but
Mb
(data
size,
megabit/byte)
is
M
=
messages
features
may
be:
2^20;
Kbps
the
K
=
10^3,
Kb
the
K
=
2^10
o Byte-‐oriented:
give
the
size
of
the
frame
11011011
R:
111
Connectivity:
Point-‐to-‐point
link
connection
Example
1:
Suppose
you
want
to
send
a
1.5
MB
(bytes)
put
the
number
of
bytes
in
the
1101
10101101000
(PRO:
security,
dedicated
(always
available)
CON:
file.
RTT
=
80
ms.
Packet
size
=
1
KB.
There
header
then
read
that
many
bytes
(don’t
1101
hard
to
scale
(add
device?
Create
a
longer
path))
is
2
x
RTT
of
handshaking.
What
is
the
total
have
to
escape
characters);
Header
–
1111
box
-‐
box
time
to
transfer
the
1.5
MB
file
in
the
Payload
-‐
Checksum
1101
Multiple
access
several
nodes
vying
for
a
local
following
cases?
a.
bandwidth
is
10
Mbps
o Bit-‐oriented:
these
protocols
beginning
and
0101
access
channel,
see
how
it
shares
link
(Ethernet)
and
packets
are
sent
continuously
ANSWER:
ending
bit
pattern
(if
this
pattern
occurs
0000
(PRO:
store
date
in
one
place
and
everyone
can
total
time
=
2
x
RTT
+
L
=
(2
x
80ms)
+
(40ms
somewhere
else
in
your
data
it
puts
a
zero
1010
th
access
it,
speed
as
its
faster
because
everyone
+(1.5MB/10Mbps)
+0)
=
.16
sec
+
(.04
sec
+
in
6
1’s
spot);
good
because
its
flexible
in
1101
can
intercept
every
message
sent,
scalable,
low
(1.5x2^20x8(bits)/10x10^b
(sits/sec))
=
that
you
can
send
variable
number
of
bits;
1111
idle
time)
line
on
bottom
and
connected
boxes
1.458
seconds
b.
bandwidth
is
10
Mbps,
but
EX:
HDLC
01111110
where
the
111111
is
the
1101
above
after
sending
each
packet,
must
wait
one
escape
pattern
0100
Scalability
RTT
before
sending
the
next
packet
First,
o Clock-‐based:
use
the
clock,
every
frame
is
0000
Sharing:
Time-‐sharing
(STDM:
Synchronous
Time
let’s
see
how
many
packets
we
need
to
exactly
the
same
size
(fixed
size
frames);
1000
Division
Multiplexing):
The
sending
nodes
share
send:
ANSWER:
#
packets
=
1.9
MB/1
KB
=
good
because
its
easy
but
bad
if
you
are
1101
the
link
by
taking
turns
via
time.
Time
is
broken
1.5x2^20B/1x2^10B
=
1536
packets;
total
sending
small
amounts
of
data
because
you
1010
into
slices
(.1
seconds
each)
so
A
gets
to
use
the
time
=
1.458
+
(1535*.08sec)
=
124.258
are
wasting
lots
of
data
room;
EX:
SONET
(9
1101
link
between
the
switches
for
.1
seconds
then
seconds
c.
bandwidth
is
infinite,
up
to
20
rows
of
90
bytes)
111
the
others,
then
A
gets
to
use
it
again
(PRO:
fair
packets
can
be
sent
each
RTT
ANSWER
From
o Sentinels
(byte
boundary):
use
a
control
P(x)
=
T(x)
–
R(x)
and
easy
to
implement)
(CON:
wasting
time
if
no
part
(b),
we
know
the
#
of
packets
is:
code
at
the
beginning
and
end
of
the
frame
data
is
waiting
to
be
sent
from
the
node
1538/20
=
76.8
=
77
transmits
but
76.5
RTT;
(means
you
need
to
escape
the
control
code
10101101000
(inefficient),
long
wait
times,
scalability
isn’t
Total
time=
(2xRTT)
+
76.5
RTT)
=
78.5
RTTs
“/”
for
example;
good
example
of
this
is
a
-‐
111
good)
=
78.5x.08sec
=
6.28sec
Example
3
(ex.
PPP
(Flag
–
Header
–
Payload
–
Checksum
–
___________
FDM:
Frequency
Division
Multiplexing:
The
1.27):
Calculate
the
bandwidth
necessary
Flag)
10101101111
signal
is
split
up
into
three
different
channels
and
for
transmitting
the
following
data
in
real
• Error
Detection
ADD
WORKSHEET
(subtraction
done
via
XOR)
each
channel
is
sent
at
a
different
frequency
time.
a.
HDTV
(1920
x
1080
pixels,
24
o 2D
parity:
each
row
and
col
has
an
even
over
the
switch-‐to-‐switch
link.
This
is
similar
to
bits/pixel,
30
frames/second)
ANSWER:
number
of
face
up
and
face
down
cards
so
10101101111
is
sent
to
receiver
how
TV
signals
work
(you
get
all
100
channels
all
data/time
=
(1920x1080x24)bits
X
30
you
can
compare
like
Sudoku
to
find
the
Receiver
calculates
message:
P(x)
/
C(x)
the
time
even
if
you
don’t
tune
into
them)
(PRO:
frames/sec
=
1.5
Gbps
exact
position;
1
bit
knows
exactly
what
bit
000
is
remainder
Thus,
the
data
is
“good”
and
all
can
talk
at
the
same
time,
multiple
senders,
Example
4
(ex.
1.17):
Calculate
the
latency
is
error;
2
bit
can
detect
but
4
possible
bits
the
message
is
10101101.
(remove
last
3
bits
fair,
good
for
real
time
data)
(CON:
sending
lots
(first
bit
sent
to
last
bit
received)
for
the
for
2
errors;
3
or
4
bits
can
detect
some
and
from
what
was
sent)
11-‐>0;
10
-‐>1;
01-‐>1;
00-‐>1
of
unnecessary
data,
inefficiently
with
idle
following
cases:
a.
1-‐Gbps
link
with
a
single
not
others
Reliable
Transmission:
Data
can
get
lost
because
nodes,
scalability)
store
and
forward
switch
in
the
path.
Packet
o Checksum:
IP
uses
that
for
each
packet;
of
buffer
overflow
on
the
switches,
corrupted
Statistical
Multiplexing:
The
data
is
broken
into
size
is
5000
bits.
Each
link
introduces
a
PRO:
no
matter
how
long
the
message
is
bits,
internal
nodes
goes
down,
sender
node
fails
packets
(with
a
fixed
maximum
size).
Packets
propagation
delay
of
10
us
and
that
the
you
are
always
sending
the
same
amount
of
ARQ:
Automatic
Repeat
Request
(at
the
link
arrive
at
the
switch
and
each
is
send
along
the
switch
begins
retransmitting
immediately
data,
which
means
scalability,
is
great.
This
layer);
sender
sends
the
frame,
wait
for
ack,
no
switch-‐to-‐switch
link.
The
switch
needs
to
after
it
has
finished
receiving
the
packet.
A
–
X
could
detect
one
flipped
bit
(cant
recover
it
ack
received
send
the
frame
again,
if
ack
arrives
determine
how
to
prioritize
the
packets.
This
–
B
and
one
link
is
one
line;
For
one
link:
like
you
can
in
2D)
but
you
would
know
an
send
next
frame:
If
ack
takes
too
long
(more
could
be
done
in
a
first-‐in-‐first-‐out
manner
or
a
Latency
=
P+T+Q
=
10ms
+
(5KB/1Gbps)
which
error
occurred.
Cant
detect
an
aligned
set
than
2
RTT)
it
might
get
sent
again;
we
just
round
robin
fashion
where
your
turn
gets
is
5ms
+
0
=
15
ms;
For
two
links:
30
ms
of
bits
with
2
switches
because
sum
of
that
toggle
b/t
frame
1
and
frame
0
because
with
skipped
if
you
have
no
data.
(PRO:
uses
the
Physical
Links
col
would
be
the
same.
ARQ
we
can
only
have
2
things
out
at
a
time;
bandwidth
well,
efficient
as
long
as
someone
has
Coaxial
cable:used
by
Comcast,
copper
wire
w/
2
Formula
for
checksum:
1.
Add
up
the
words
PRO:
simple
and
it
gets
you
what
you
want
-‐>
something
to
send
the
switch
is
working,
layers
of
insulation;
10
-‐100
Mbps
can
go
200
m
transmitted
using
ones
complement
arithmetic.
you
know
when
someone
gets
your
message;
scalability,
Internet
uses)
(CON:
wait
times
could
Twisted
pair:
like
Ethernet
has
4
twisted
pairs,
2.
Use
the
sum
as
the
checksum
(represented
in
CON:
sender
waits
a
lot
so
not
utilizing
the
increase
a
lot,
less
fair
than
other
2,
more
work
CenturyLink,
10
–
100
Mbps
can
go
100
m
ones
complement)
bandwidth
which
makes
it
very
slow
for
nodes
and
switches)
Fiber:
made
of
glass
and
sends
lights
as
signals;
Review
of
ones
complement
representation:
Stop
&
Wait
Network
architecture
&
layers
100
Mbps
–
10
Gbps
can
go
40
km
Positive
numbers:
represented
normal
in
binary
Sliding
Window:
PRO:
not
nearly
as
much
wait
OSI
(7
layers):
Application
(whatever
application
• Encoding
Schemes
–
lots
of
transitions
make
Negative
numbers:
each
bit
is
inverted
Examples
time,
utilizing
bandwidth;
CON:
more
requires),
Presentation
(encodes/decodes
data,
the
signals
easy
to
decode
of
ones
complement:
7
=
00000111,
-‐7
=
complicated
to
code,
more
things
to
keep
track
formatting),
Session
(management
of
streams),
o NRZ:
1
-‐>
high,
0
-‐>
low;
PRO:
simple
to
11111000
Assume
data
is
the
following:
of
more
things
to
process
HANDOUT
Transport
(reliability,
acks
and
congestion
implement
and
simple
to
understand;
CON:
00000010(2)
00000101(5)
00000011(3)
LANs
control),
Network
(route
packets),
Data
Link
long
repeated
bits
(leads
to
baseline
11111100(-‐3)
Then,
when
we
add
the
numbers,
MAC
addresses
(hardware
addresses)
(framing,
encoding
bits
into
signals),
Physical
wander)
so
that
if
you
have
a
lot
of
zeros
we
get:
00000010(2)
00000101(5)
=
00000111
Aloha
–
early
version
of
using
a
shared
resource:
(actually
sends
signal
across
the
link)
the
signal
will
wander
to
the
average
and
(7)
+
00000011(3)=
00001010(10)
+
11111100 Hawaiian
islands
if
you
had
data
to
send,
you
Internet
(4
layers):
1
-‐
Application
(Application,
make
it
harder
to
decode.
(-‐3)
=
00000110
(plus
a
1
in
the
carry
sent
it
and
if
you
collided
then
you
picked
a
Presentation,
Session)
(EX:
HTTP
(web
fetches),
o NRZI:
1
-‐>
transition,
0
-‐>
no
transition;
PRO:
bit,
so
this
is
added
to
the
LSB)+1
=
00000111(7)
random
length
of
time
and
tried
to
send
it
again;
SMTP
(email),
FTP
(files
from
one
place
to
long
sequence
of
1’s
we
are
now
okay;
CON:
So,
00000111
would
be
sent
as
the
checksum
for
P
persistent
wait
a
random
amount
of
time
another))
2
–
Transport
(Transport)
(EX:
TCT
still
have
baseline
wander
for
zero’s
the
data.
Ethernet
(802.3)
(reliable
data,
most
used),
UDP
(video,
no
acks
o Manchester:
1
-‐>
high
to
low,
0
-‐>
low
to
o CRC:
Pro:
Low
overhead
(only
going
to
add
CSMA/CD
(Carrier
Sense
(a
node
can
detect
built
in,
good
for
high
volume
data))
3
–
Network
high;
PRO:
lots
of
transitions;
CON:
has
to
the
number
of
bits
in
C(x)
which
is
an
whether
a
line
is
idle
or
busy)
Multiple
Access
(Network)
(EX:
IP)
4
–
Link
(Link,
Physical)
(EX:
read
w
signals
for
each
bit
so
only
running
at
advantage
over
checksum,
its
more
unlikely
(shared
common
resources)
with
Collision
Ethernet,
wireless,
fiber,
cell
protocols,
etc.)
50%
efficiency
so
we
are
loosing
half
the
that
if
you
flip
a
couple
of
bits
you
will
still
Detection
(node
listens
while
sending
and
can
Network
Components
amount
of
data
we
could
be
sending
get
a
correct
remainder
of
0.
CONS:
more
detect
collisions))
o Nodes
(connection
point),
links,
hosts
o Differential
Manchester
-‐
the
receiver
does
work
(for
both
sender
and
receiver)
not
a
lot
Protocol
for
sending:
1.
Send
if
you
have
data
2.
(COMPUTERS
that
host
information),
not
need
to
know
the
polarity
of
the
signal.
but
a
litter
and
there
is
no
error
recovery.
If
link
is
busy
wait
and
then
send;
this
is
1
switches
(bridges),
routers
(gateways)
The
polarity
can
be
figured
from
the
line
Internets
idea
is
to
know
if
something
went
Persistent
or
it
sends
ASAP
and
if
there
is
a
Network
Data
transitions.
It
does
not
matter
whether
a
wrong
but
we
don’t
really
care
to
fix
it
–
just
collision
both
have
to
wait
o Frames(link
layer),
packets
(network
layer)
logical
1
or
0
is
received,
but
only
whether
have
sender
resend
data
Each
frame
has
a
min
size
so
they’re
are
long
messages
(IP
header,
checksum,
etc.,
)
(all
the
polarity
is
the
same
or
is
different
from
Background:
Represent
message
M
by
a
enough
to
detect
collisions
(64
B)
and
a
max
size
the
same
–
just
different
names
on
different
the
previous
value.
polynomial
M(x)
If
M
=
10010101,
then
M(x)
=
x7
so
the
frame
isn’t
too
big
because
then
we
allow
layers)
o 4B/5B
–
encode
a
4
bit
pattern
into
5
bits
so
+
x4
+
x2
+
x0
one
to
hog
and
increase
the
chance
of
a
waiting
• FDM,
STDM,
statistical
multiplexing
you
can
get
rid
of
a
lot
of
zeros
in
a
row
Sender
calculates
P(x)
from
M(x)
that
is
exactly
queue
(1500
B)
(protocols
for
multiplexing
flows)
then
apply
NRZI;
PRO:
Transitions;
CON:
divisible
by
C(x),
where
C(x)
is
an
established
Coax
can
be
500
m
away
and
a
twisted
pair/fiber
Network
Performance
80%
efficiency
agreed
upon
polynomial
with
degree
k.
Send
can
be
2500
m
away
but
we
can
make
this
bigger
o
Latency:
Total
time
for
a
transfer,
also
P(x).
by
adding
switches
called
delay,
L
=
P
+
T
+
Q
where
P
is
the
Receiver
divides
P(x)
by
C(x).
If
remainder
is
not
Frame
Format:
Preamble
(8B,
alternating
0’s
and
propagation
delay
and
the
physical
time
to
0,
an
error
occurred.
M(x)
is
all
but
the
last
k
bits
1’s
to
sync
reception)
–
Dest
(6B)
–
Source
(6B)
–
transfer,
depends
on
the
type
of
wire
and
and
is
used
as
the
message
To
calculate
P(x):
Type
(2B
what
higher
level
protocols)
–
Data/Payload
–
CRC
(4B,
error
detection)
Ethernet
everyone
hears
everything
so
everyone
Idea
of
algorithm:
1.
Elect
root
node
(use
lowest
640x480x3x30bytes/second
b.
Song
with
length
hears
the
destination
address,
is
it
for
you?
address)
2.
Grow
tree
as
shortest
distances
from
of
2
minutes
and
total
size
of
30
MB:
30MB/2
Accept.
If
not
ignore
the
root
bridge.
Break
ties
with
lowest
address.
min
=
30x2^20/2
min
=
.26
Mbps
If
destination
is
6
1’s
then
it’s
a
broadcast
so
Bridges
send
config
messages
over
ports
for
HOMEWORK
2
everyone
should
receive.
1
followed
by
5bit
which
they
are
the
best
path.
Turn
off
ports
that
4.
Consider
n
ARQ
protocol
running
over
a
20-‐km
group
ID
is
a
multicast,
if
you
are
a
member
of
are
not
on
best
paths
3.
LAN
uses
its
designated
point-‐to-‐point
fiber
link.
A.
Calculate
the
the
group
you
should
listen.
bridge
(one
with
port
still
active)
and
the
propagation
delay
for
this
link,
assuming
the
Collisions:
use
exponential
back
off
(1st
collisions
Dest
(Port)
designated
bridges
do
the
forwarding
across
speed
of
light
is
2x
10^8
m/s
in
fiber:
Prop
Delay
wait
0
or
1
frame
times,
2nd
collision
wait
0,
1,
A
(X1)
B
(X1)
C
(C)
D
(D)
E
(X3)
F
(X3)
G
(X3)
LANs
Algorithm:
1.
Each
bridge
believes
it
is
the
=
(20x10^3
m/2x10^8
m/s)
=
.1ms
b.
Suggest
a
2,
or
3
frame
times)
cap
N
at
10
not
the
most
fair
root.
When
learn
not
the
root,
stop
sending
suitable
timeout
for
the
ARQ
protocol
to
use
for
but
works
relatively
well
in
lightly
loaded
CON:
limited
table
size/tables
too
large/tables
config
(hello)
messages.
Forward
root’s
config
this
link:
A
suitable
ARQ
protocol
timeout
would
networks.
can
get
out
of
date,
don’t
know
anything
about
message
with
#
hops
incremented
by
1.
Records
be
2
RTTs
to
send
data
back
and
forth
and
forth
Max
number
of
nodes
on
a
network
is
1024
but
the
nodes
your
sending
data
too
best
config
for
each
port
2.
When
not
a
and
back
c.
why
might
it
still
be
possible
for
the
today
more
common
that
every
node
is
PRO:
simple,
we
don’t
have
to
know
anything
but
designated
bridge,
stop
forwarding
config
ARQ
protocol
to
time
out
and
retransmit
a
connected
to
a
switch.
PRO:
easy
to
administer
a
table
lookup,
we
can
figure
out
it
packets
messages
3.
Real
root
sends
config
messages
frame,
given
this
timeout
value?
It
might
not
(EASY
to
add
another
node)
CON:
Doesn’t
work
aren’t
getting
there
and
we
can
change
periodically
4.
If
bridge
does
not
receive
a
config
work
if
the
node
has
to
process
lots
of
data
or
it
very
well
on
a
crowded
network
forwarding
table
route
thorough
a
new
message
in
a
certain
period
of
time,
assume
processes
data
slowly
in
general.5.
Remember
You
can
mitigate
collisions
by
limiting
the
intermediate
switch
topology
has
changed
and
start
sending
config
the
timeline
diagrams
for
the
sliding
window
number
of
nodes,
the
distance
so
RTT
delay
is
Virtual
Circuit
Switching
messages
claiming
to
be
the
root
Example:
algorithm
with
SWS
=
3
frames
and
timeout
much
less
than
the
max).
More
complicated
for
the
switch.
Each
switch
Assume
no
bridge
has
any
info
about
any
other
interval…lines
right
and
left
6.
Wireless
(Wi
Fi
802.11)
has
a
table
with:
Incoming
Port,
VCI
in,
Outgoing
bridge
in
network.
Let’s
look
at
B3:
1.
B3
sends
CSMA/CA
(Carrier
Sense
Multiple
Access
with
Port,
VCI
out.
Every
switch
knows
exactly
how
(B3,
0,
B3)
to
B5
and
B2
[claiming
to
be
root]
2.
Collision
Avoidance)
many
connections
are
going
through
it
all
the
B3
receives
(B2,
0,
B2)
and
(B5,
0,
B2)
from
B2
PAN
(Bluetooth)
LAN
(WiFi)
and
WAN
(3G)
time
because
=
the
number
of
table
entries
(so
and
B5,
respectively.
Since
B2
is
<
B3,
B3
accepts
Signal
degrades
as
distance
increases.
we
know
the
load
at
all
times).
Table
has
more
B2
as
root
3.
B3
sends
(B2,
1,
B3)
to
B5
to
Wireless
cant
hear
everyone
else
like
you
can
on
entries
because
different
connections
could
be
forward
message
[note
that
the
#hops
is
Ethernet
because
of
hidden
nodes
problem
so
sending
simultaneously.
The
outgoing
ID
is
the
incremented]
4.
B3
receives
(B1,
1,
B2)
from
B2
instead
wireless
listens
and
sees
if
any
signals
are
VC
for
the
next
hop.
Sequence
of
Events
for
and
(B1,
1,
B5)
from
B5.
Since
B1
<
B2,
B3
being
propagated
within
our
area,
if
so
wait,
if
Sending:
When
A
wants
to
send
to
F:
1. accepts
B1
as
root.
5.
B3
could
send
(B1,
2,
B3)
to
idle
you
reserve
the
medium
or
signal.
If
signals
Put
the
ID
as
5
in
header
of
packets
B2
or
B5,
but
it
does
not
since
it
is
nowhere
the
collide
within
your
radius
you
still
do
the
destined
to
F
Then,
S1
gets
packet:
1.
Looks
up
“shortest
path”
from
B1.
So,
B3
is
not
a
exponential
back
off
values
5
and
A
in
table
2.
Removes
5
as
the
VC
designated
bridge.
6.
B3
receives
(B1,
1,
B2)
from
Say
B
wants
to
send
to
C
and
A
wants
to
send
to
ID
in
header
3.
Puts
8
as
the
VC
ID
in
header
4.
B2
and
(B1,
1,
B5)
from
B5
again,
so
network
is
B.
Band
C
within
range
and
same
with
A
and
B
Forwards
packet
along
outgoing
port
X2.
How
do
stable.
B3
turns
off
data
forwarding
to
LANs
A
(each
circle
covers
3).
Events:
1.
B
sends
C
a
RTS
VCIs
get
assigned?
Signal:
A
sends
a
startup
and
C.
frame
with
duration
(reservation)
2.
A
hears
RTS
message
to
switch
1
with
destination
F.
Switch
1
HOMEWORK
1
and
defers
3.
C
responds
to
B
with
a
clear
to
forwards
message
to
switch
2
with
an
unused
2.
Calculate
the
total
time
required
to
transfer
a
send
(CTS)
frame
4.
D
hears
CTS
and
defers
5.
B
VCI,
fill
in
incoming
and
outgoing
port
entries.
1000
KB
file
in
the
following
cases.
You
may
can
send
data
to
C
6.
C
sends
acks
back
to
B
assume
a
RTT
of
100
ms,
a
packet
size
with
1
KB
10.
Give
the
circuit
table
for
all
the
switches
after
Switch
2
forwards
message
to
switch
3
with
an
each
of
the
following
connections.
Host
A
If
a
node
hears
a
RTS
and
not
a
CTS
then
it
can
unused
VCI,
fill
in
ports.
Switch
3
forwards
of
data,
and
an
initial
2
x
RTT
of
“handshaking”
send
because
not
sending
to
same
circle.
before
data
is
sent.
a.
The
bandwidth
is
1.5
Mbps
connects
to
host
C,
Host
D
connects
to
Host
G,
message
to
F.
F
chooses
a
VCI.
F
sends
ack
to
Host
E
connects
to
host
I,
Host
D
connects
to
Access
points
send
out
beacon
frames
and
that’s
Switch
3
with
VCI,
so
switch
3
uses
it
for
its
and
data
packets
are
sent
continuously:
2
initial
what
computers
hears
RTT
+
transmit+
P
delay
+
Q
delay=
200
ms
Host
B,
Host
F
connects
to
Host
I,
Host
H
outgoing
ID.
Switch
3
forwards
ack
with
the
VCI
it
connects
to
host
A
Connecting
to
AP:
1.
Node
sends
a
probe
frame
used
in
table.….
Ack
gets
back
to
A
and
a
virtual
+(1000KB/1.5Mbps)+50ms
and
middle
term
Switch
In
Port
VCI
in
Outgoing
VCI
out
2.
AP
requires
with
a
probe
response
frame
3.
connection
is
created.
looks
like
(1000x2^10x8for
bits)/1.5x10^6
=
1
A
0
C
1
Node
selects
and
AP
send
Association
request
PRO:
telephone
industry
uses.
Know
load
and
5.4613
seconds
and
back
to
ordinal
=
.20
seconds
1
D
2
2
3
Frame
4.
AP
Associate
Response
more
reliable.
Faster
connections
established.
+(.05+5.4613)
=
5.7113
seconds
b.
The
2
2
3
4
0
4
4
0
G
1
Frame:
Control
(16B,
RTS,
STS,
if
AP’s
other
bits
Cycles
are
allowed
(redundancy
resilience)
bandwidth
is
1.5
Mbps
but
after
sending
each
2
E
1
2
2
for
management)
–
Duration
(16B,
length
based
CON:
every
switch
still
needs
a
forwarding
table
data
packet,
we
must
wait
one
RTT
before
2
2
2
3
5
on
sender
data
size
and
rate)
–
Addr1
–
Addr2
–
and
you
have
to
set
up
and
tear
down
all
the
sending
the
next:
We
have
100
packets
-‐>
999
3
3
5
I
0
1
D
4
1
5
Addr3
(48B
each,
with
4
they
make
up
the
MAC
connections
(table
entries)
wait
times
so
same
as
part
a
+
wait
times
=
2
1
5
2
6
addresses)
–
Sequence
Control
(16B
frame
#)
–
Source
Routing
5.7113
+(.1)(999)
=
105.6113
seconds
c.
The
3
2
6
3
1
Addr4
–
Payload
(0
–
18996
B)
–
CRC
(32B)
Most
difficult
for
the
node.
Instead
of
switches
bandwidth
is
“infinite”
(transmit
time
is
0)
and
3
3
1
B
2
Bluetooth:
Uses
a
master-‐slave
model
that
up
to
20
packets
can
be
sent
per
RTT:
100/20
=
2
F
7
2
8
dong
lots
of
work
the
end
nodes
do
all
the
work.
2
2
8
3
9
makes
it
very
simple.
Slaves
can
only
Nodes
have
to
know
topology
of
entire
network.
50
total
but
49.5
RTTs;
Total
time
=
(2xRTT)
+
3
3
9
I
4
communicate
the
with
the
baser
and
you
can
If
A
sends
to
F:
In
header,
A
puts
the
sequence
F
49.5
RTT
=
51.5
RTTs
=
51.5(.1)
=
5.15
seconds
4
H
2
2
10
have
up
to
7
slaves.
Time
is
broken
into
time
–
X3
–
X2.
Then,
each
node
looks
at
last
address
3.
Suppose
a
100
Mbps
point-‐to-‐point
link
is
2
2
10
1
11
1
1
11
A
6
plots
if
time
slots
are
numbered
a
frame
could
for
forwarding
and
rotates
the
entry
order
being
set
up
between
Earth
and
the
moon.
The
take
up
1,
3,
or
5
slots.
Slaves
can
only
send
distance
between
the
Earth
and
the
moon
is
(moves
last
to
first).
So,
then
X1
would
see
the
during
even
time
slots
and
master
odd.
Slaves
header,
rotate
it
to
X2
–
F
–
X3
and
send
it
to
X2.
385,000
km
and
the
data
travels
at
the
speed
of
only
respond
too
the
master
and
Bluetooth
uses
PRO:
Easy
for
switches
light
–
3
x
108
meters
per
second.
a.
Calculate
frequency
hopping.
CON:
nodes
have
to
have
a
mechanism
to
get
the
minimum
RTT
for
the
link:
RTT
=
2xP
=
2x
3G
has
3
options.
1)
TDMA:
Everyone
gets
a
time
topology
info
(385000Km/3x10^8
m/s)
=
2x(385000x10^3
slot,
bandwidth
and
thru
put
isn’t
good.
2)
Learning
sources
/
ports
m/3x10^8
m/s)
=
2.57
seconds
b.
Using
the
FDMA:
combine
all
data
and
all
send
at
different
LAN
connections
listen
all
the
time
to
everyone’s
minimum
RTT
as
the
delay,
calculate
the
delay
x
frequency
(like
TV)
bandwidth
again
isn’t
great
3)
data.
We
only
add
to
the
table
by
source,
don’t
bandwidth
product
for
this
link.
DxB
=
100
Mbps
CDMA:
We
all
talk
at
once
but
we
all
talk
in
a
X
(2.57seconds/2)
=
1.28x10^8
bits
c.
A
camera
care
about
the
destination.
Constantly
listening
different
language
and
the
cell
tower
can
speak
to
know
all
source
information
and
each
entry
on
the
moon
takes
pictures
of
the
Earth
and
9.
Give
the
tables
after
A-‐>
C,
C-‐>A,
D
-‐>
C
all
languages
so
everyone
gets
a
special
code
and
lives
in
our
table
for
a
fixed
amount
of
time.
saves
them
in
digital
form.
Assume
you
want
to
you
XOR
your
data
before
sending
(data
retrieve
a
25
MB
image
from
the
moon.
What
is
Switch
Host
Port
Cycles
may
send
packets
around
indefinitely.
So
encryption)
and
all
signals
are
added
together
the
minimum
amount
of
time
that
elapses
B1
A
A
we
build
STs.
See
HW2.
and
then
receiver
XOR
data
backwards
to
know
between
the
initial
request
from
Earth
and
the
B1
C
B2
o Spanning
Tree
Algorithm
(Perlman
paper)
who
is
was
sent
from.
The
built
in
error
entire
image
is
transferred
back
to
Earth.
1
RTT
+
B2
A
B1
If
a
bridge
fails
the
connected
bridges
stop
detection
scheme
is
a
real
benefit.
Just
ignores
getting
config
messages
so
they
send
out
the
I’m
T
=
2.57
seconds
+
25
MB/100Mbps
=
2.57
+
B2
C
B3
everything
it
hasn’t
given
a
code
out
for
so
it
has
the
root
message
and
we
rebuild
the
ST.
25x2^20x8
=
4.57
seconds
B2
D
B4
built
in
interference
security.
4.
Calculate
the
latency
(from
first
bit
sent
to
last
B3
A
B2
Switches
(Bridges)
and
Forwarding
–
helps
with
bit
received)
for
the
following:
a.
A
10
Mbps
B3
C
C
scalability
Ethernet
with
a
single
store-‐and-‐forward
switch
B3
D
B2
Datagram:
what
networks
use
today;
needs
a
in
the
path
and
a
packet
size
of
5000
bits.
B4
A
B2
buffer
because
it
takes
data
in
and
sends
it
out
Assume
that
each
link
introduces
a
propagation
B4
E
B2
on
the
correct
cable.
Also
gives
us
a
way
to
delay
of
10
us
and
that
the
switch
begins
B4
D
D
connect
data
from
one
LAN
to
another
and
retransmitting
immediately
after
it
has
finished
switch
deals
with
knowing
where
to
forward
the
receiving
the
packet:
L
=
P+T+Q=
10
ms
data
out
too
and
on
what
line.
We
label
the
+5Kb/10Mbps
=
.01
sec
+
(5x2^10b/10x10^6
b/s)
ports
by
the
direct
link
they
are
connected
too.
=
14.1
ms
per
link
=
28.2
ms
b.
Same
as
(a)
but
Forwarding
tables
only
know
the
next
hop
and
with
3
switches
along
the
path:
56.4
ms
every
switch
ahs
its
own
forwarding
table
5.
For
each
of
the
following
situations,
calculate
the
bandwidth
necessary
to
transmit
the
data
in
real
time
(assume
no
compression
is
done).
a.
Video
at
a
resolution
of
640
x
480
pixels,
3
bytes
per
pixel,
30
frames/second: