to download the slides of the sessions on IP Technology

cursefarmΔίκτυα και Επικοινωνίες

24 Οκτ 2013 (πριν από 3 χρόνια και 11 μήνες)

79 εμφανίσεις

IP Technology
Introduction
IP Address Classes
IP
Subnetting
CIDR
Routing Basics
IP protocol
ICMP
ARP
DHCP
NAT
Virtual
-
Circuit vs.
Datagram Subnets
5
-
4
QoS Requirements of Typical Apps
Internetworking
Interconnecting LANs by Hub, Switch, or Router?
Concatenated Virtual Circuits
Connectionless Internetworking
A connectionless internet.
IP Addresses
IP address formats.
IP
Special Addresses
Network
0000
. . .
0000
Network address
Subnetting

A campus network consisting of LANs for various
departments
Subnetting consists of assigning a “range” of addresses to each subnet. The subnets
need not be of the same size, but their sizes must be “powers of
2
”, so that they have
non overlapping subnet addresses and an integral number of bits for host numbering.
Subnetting
example

Divide the network
192
.
168
.
50
.
0
/
24
into
4
equal subnets
The easiest way is to work in decimal form and then write the final results in
binary. The initial address range is
192
.
168
.
50
.
0
to
192
.
168
.
50
.
255
. We can
divide it to many ranges but the only condition is that these ranges have to
be
powers of
2
.
The technique is slightly more difficult when the subnets are not of equal size
and the initial network not a class
-
A, B or C but an arbitrary subnet.
However the method is strictly the same, it only takes a little more time.
Subnet
Range
address
mask
A
192
.
168
.
50
.
0
to
63
192
.
168
.
50
.
0
/
26
255
.
255
.
255
.
192
B
192
.
168
.
50
.
64
to
127
192
.
168
.
50
.
64
/
26
255
.
255
.
255
.
192
C
192
.
168
.
50
.
128
to
191
192
.
168
.
50
.
128
/
26
255
.
255
.
255
.
192
D
192
.
168
.
50
.
192
to
255
192
.
168
.
50
.
192
/
26
255
.
255
.
255
.
192
Subnetting
Example (
2
)
A company is assigned the IP network address
154
.
128
.
10
.
0
/
23
. It is divided into
6
departments as indicated in the following table. The second column gives for
each department the number of hosts currently available and must be assigned
public IP addresses.
Dept
#hosts
Range
IP

CIDR
maxHosts
Head office
128
154
.
128
.
10
.
0
154
.
128
.
10
.
255
154
.
128
.
10
.
0
/
24
254
Engineering
100
154
.
128
.
11
.
0
154
.
128
.
11
.
127
154
.
128
.
11
.
0
/
25
126
Marketing
20
154
.
128
.
11
.
128
154
.
128
.
11
.
159
154
.
128
.
11
.
128
/
27
30
Sales
25
154
.
128
.
11
.
160
154
.
128
.
11
.
191
154
.
128
.
11
.
160
/
27
30
Research
25
154
.
128
.
11
.
192
154
.
128
.
11
.
223
154
.
128
.
11
.
192
/
27
30
Customer Care
5
154
.
128
.
11
.
224
154
.
128
.
11
.
231
154
.
128
.
11
.
224
/
29
6
Available
154
.
128
.
11
.
232
154
.
128
.
11
.
255
CIDR

Classless
Inter
-
Domain
Routing
The idea of subnetting has been generalized as a flexible way to
assign addresses on the Internet (due to shortage of addresses).
This scheme is better than class
-
based addressing for managing
the shortage of IP addresses
CIDR
How to generalize the idea of
subnetting
to Classless Inter
-
Domain routing?
Routers introduce the subnet mask in their routing tables:
Subnet
Mask
NextHop
194
.
132
.
101
.
0
255
.
255
.
255
.
0
to the main router of the
univ
.
p.s. in this example we can maintain only one entry thanks to “
path aggregation

CIDR
What if a subnet moves to another ISP?
R
3
’s routing table should look like this

An entry matches a packet destination
ifif
:
Dest
& Mask == Subnet
temporarily R
3
’s RT might look like this

Therefore, more than one entry might
match a destination =>
longest match
Subnet
Mask
NextHop
194
.
132
.
101
.
128 255
.
255
.
255
.
224
R
4
194
.
132
.
101
.
0 255
.
255
.
255
.
224
R
2
194
.
132
.
101
.
32 255
.
255
.
255
.
224
R
2
. . .
194
.
132
.
101
.
224 255
.
255
.
255
.
224
R
2
Subnet
Mask
NextHop
194
.
132
.
101
.
128 255
.
255
.
255
.
224
R
4
194
.
132
.
101
.
0 255
.
255
.
255
.
0
R
2
Host configuration
An IP host has at least one IP per network interface.
Each interface has the following configuration parameters:
IP, subnet mask, default gateway, DNS server(s)
The default gateway must be on the same subnet.
Q: When does a host forward a packet to the default gateway?
A: when the destination address is not on the same subnet of any of its
interfaces
Q: how does a host know if a
destIP
address is on the same subnet?
A: if(
myIP
&
myMask
==
destIP
&
myMask
) then it’s a neighbor.
Q: what does a host do if the
destIP
is on the same subnet?
A: deliver the packet directly over the physical network’s protocol. The
physical address of the destination host is resolved from its
destIP
using
the
ARP
protocol.
ARP

The Address Resolution Protocol
The IP
Packet Format (V
4
)
Fragmentation
(a)
Transparent fragmentation.
(b)
Nontransparent fragmentation.
NAT

Network Address Translation
Placement and operation of a NAT box.
Port=
7515
Port=
5640
ICMP
-
Internet
Control Message Protocol
5
-
61
DHCP
-
Dynamic
Host Configuration Protocol
Transport layer

Basic transport
running in the kernel

UDP:
datagram service

TCP:
connected service

New: Multipath TCP (MPTCP)

Additional transport mechanisms
running in the app layer

RTP (/UDP)
multimedia Realtime

HTTP (/TCP)
Web Applications

RPC (/UDP)
distributed programming, middleware..

ORBs (/UDP)
OO distributed programming

Web
-
Services (/HTTP)

MQ (/TCP)
offline reliable messaging (banking apps)
UDP

datagram
-
oriented transport
The UDP header
TCP:
The connected, reliable
T
ransport
C
ontrol
P
rotocol
TCP Header
Connection Establishment
-
3
way handshake
The old duplicate problem:
ISNs x and y are generated in a way to detect and reject any old
duplicates with the same (IP
1
, port
1
, IP
2
, port
2
)
Goal: use always ISN > last seq. But to avoid high memory demands,
the algorithm generates ISN from a clock that runs faster than any
sequence rate in any connection. Also, seq numbers should not wrap
Some assigned ports
Port
Protocol
Use
21
FTP
File transfer
23
Telnet
Remote login
25
SMTP
E
-
mail
69
TFTP
Trivial File Transfer Protocol
79
Finger
Lookup info about a user
80
HTTP
World Wide Web
110
POP
-
3
Remote e
-
mail access
119
NNTP
USENET news
NAT

Network Address Translation
Placement and operation of a NAT box.
Port=
7515
Port=
5640
The Sockets API
Socket API Example
File transfer
The Client Part
Socket API Example
File transfer
The Server Part
TCP Transmission Policy
Window management in TCP.
TCP Transmission Policy
Silly window syndrome…
Solutions: avoid sending one
-
byte segments (Nagel) and avoid
sending one byte window updates (Clark)
TCP Timer Management
(a)
Probability density of ACK arrival times in the data link layer.
(b)
Probability density of ACK arrival times for TCP.
TCP Timing is based on estimation of average and deviation of RTT
TCP Timer Management

TCP Timing is based on estimation of the average and absolute
deviation of RTT.. After each sample, M, of the round
-
trip time,
it changes its estimation according to the “smoothed average”
formula:
RTT =
α
RTT + (
1

α
) M
(average estimation)
D =
α
D + (
1

α
) | M

RTT |
(the absolute deviation)

Typically
α
=
7
/
8
.

The Acknowledgment timeout is then set to
RTT +
4
D
TCP Congestion Control
(a)
Fast network feeding low capacity receiver => flow control
(b)
Slow network feeding high
-
capacity receiver => congestion
control
TCP Congestion Control
An example of the Internet congestion algorithm.
The Real
-
Time Transport Protocol
The RTP header.
DNS

The Domain Name system
A portion of the Internet domain name space.
Resource Records
The principal DNS resource records types.
Resource Records
A portion of a possible DNS database for
cs.vu.nl.
Name Servers
Part of the DNS name space showing the division into zones.
Name Servers
How a resolver looks up a remote name in eight steps.
This is the recursive method. Alternative is the non
-
recursive
method