Download COS 338_day15

COS 338
Day 15
DAY 15 Agenda

Capstone Proposal Overdue


Capstone progress reports still overdue


3 accepted, 3 in mediation
I forgot to mark in calendar so I will grant a reprieve
Lab 4 write-up corrected

2 A’s, 1 B, 2 F’s and 1 non-submit
 Again grades are determined by effort

Lab 5 Due November 3

Assignment 5 Posted



Due November 7
Should be on assignment 7 (I suspect that we will have only 8-9 assignments)
Today we will discussing TCP/IP

Lab 6 on Thursday
2
TCP/IP
Internetworking
Chapter 8
Panko’s
Business Data Networks and Telecommunications, 5th edition
Copyright 2005 Prentice-Hall
Perspective

Chapters 4 & 5 covered single LANs

Chapter 7 covered single WANs

Most corporations have intranets that combine
multiple LANs and WANs

Most intranets use TCP/IP standards

So does the global Internet

Chapter 8 deals with TCP/IP internetworking
4
5
Internetworking with Routers
Routers Connect Multiple Networks
(LANs and WANs) into an Internet
Site B
Router
Z
LAN 1
LAN 4
Router W
LAN 3
LAN 2
Router
X
WAN
Router Y
Site A
Site C
6
Figure 8-1: Major TCP/IP Standards
User Applications
5 Application
HTTP
4 Transport
3 Internet
SMTP
Many
Others
TCP
Supervisory Applications
DNS
Routing Many
Protocols Others
UDP
IP
ICMP
ARP
2 Data Link
None: Use OSI Standards
Internetworking is done at the
internetNone:
and transport
1 Physical
Use OSIlayers.
Standards
There are only a few standards at these layers.
Note: Shaded protocols are discussed in this chapter.
7
Figure 8-1: Major TCP/IP Standards, Continued
User Applications
5 Application
HTTP
4 Transport
SMTP
TCP
Many
Others
Supervisory Applications
DNS
Routing Many
Protocols Others
UDP
At the application layer, there are
3 Internet user applications and
IPsupervisory applications.
ICMP
ARP
We will look at two supervisory applications in this chapter.
2 Data Link
None: Use OSI Standards
1 Physical
None: Use OSI Standards
Note: Shaded protocols are discussed in this chapter.
Figure 8-2: Recap: IP, TCP, and UDP
Layer
Protocol
ConnectionOriented?
Reliable? Lightweight or
Heavyweight?
4 (Transport)
TCP
Yes
Yes
Heavyweight
4 (Transport)
UDP
No
No
Lightweight
3 (Internet)
IP
No
No
Lightweight
8
9
Figure 8-3: Hierarchical IP Address
Network Part (not always 16 bits)
Subnet Part (not always 8 bits)
Host Part (not always 8 bits)
Total always is 32 bits
128.171.17.13
The Internet
UH Network (128.171)
32-bit host
IP addresses
have three parts
CBA Subnet (17)
Host (13)
10
Figure 8-3: Hierarchical IP Address, Continued

Question.

The IP address is 123.16.22.47

How large is the network part?
Figure 8-4: Border Router, Internal Router,
Networks, and Subnets
Border
Router
Internal
Router
Corporate
Network
192.168.x.x
ISP
Network
60.x.x.x
Border Routers Connect Different Networks
11
Figure 8-4: Border Router, Internal Router,
Networks, and Subnets, Continued
Subnet 192.168.2.x
Internal
Router
Subnet 192.168.3.x
Subnet
192.168.1.x
Border
Router
Corporate Network
192.168.x.x
Internal Routers Connect Different Subnets within the Firm
12
13
Figure 8-5: Part of an Internet
Router A
Router B
Router C
Router B connects to 4 subnets via its 4 interfaces (ports)
Subnet
172.30.20.x
Ethernet
Switch 1
Router D
172.30.20.
1
C1-…
Client PC R
172.30.20.
47
A1-…
Ethernet
Switch 2
Server X
172.30.20.
19
B1-…
Server Y
172.30.21.
86
D1-…
Router E
172.30.21.
2
E1-…
Subnet
172.30.21.x
Router F
172.30.21.
1
F1-…
Figure 8-5: Part of an Internet, Continued
Router A
Interface 4
172.30.19.1
11-…
Subnet
172.30.20.x
Router D
172.30.20.
1
C1-…
Router B
Router B
Interface 1
Subnet 172.30.19.x
802.11
14
Router C
Router B’s Interface 1
is connected to a point-to-point 802.11 subnet,
Subnet
Ethernet
172.30.19.x
Ethernet
Switch 1
Switch 2
172.30.21.x
This subnet goes to Router A’s Interface 4,
Client PC R which
Serverhas
X IP address 172.30.19.1
and MAC address
172.30.20.
172.30.20.
Server Y11- …
Router E
47
19
172.30.21.
172.30.21.
A1-… Each interface
B1-…
on a router
86 has a different
2
E1-…
IP address and dataD1-…
link layer address.
Router F
172.30.21.
1
F1-…
Figure 8-5: Part of an Internet, Continued
Router A
Subnet
172.30.20.x
Router D
172.30.20.
1
C1-…
Router B
15
Router B
Router C
Interface 4
Subnet 172.30.22.x
802.11
Interface 1
172.30.22.9
21-…
Router B’s interface 4 also connects
To an 802.11 point-to-point subnet,Ethernet
Ethernet
172.30.22.x.
Switch 2
Switch 1
This reaches Interface 1 on Router C.
Client PC R
Server X
This
interface has
172.30.20.
172.30.20.
Server Y
Router E
IP address 172.30.22.9
47
19
172.30.21.
172.30.21.
and
MAC
address
21…
A1-…
B1-…
86
D1-…
2
E1-…
Subnet
172.30.21.x
Router F
172.30.21.
1
F1-…
16
Figure 8-5: Part of an Internet, Continued
Router A
Router B
Router B
Interface 2
Ethernet
Subnet
172.30.20.x
Ethernet
Switch 1
Router D
172.30.20.
1
C1-…
Client PC R
172.30.20.
47
A1-…
Router C
Router B’s Interface 2
connects to Ethernet subnet
172.30.20.x.
Subnet
Ethernet
This subnetSwitch
has a 2single172.30.21.x
switch.
Server X
172.30.20.
19
B1-…
Other devices on the subnet
include a single router (D),
Server Y
Router E
Router F
a
single
Client
PC
(R),
172.30.21.
172.30.21.
172.30.21.
and
(X). 1
86 a single server
2
D1-…
E1-…
F1-…
Figure 8-5: Part of an Internet, Continued
Router A
Router B
Router B’s Interface 3
connects to Ethernet Subnet
172.30.21.x
17
Router C
Router B Interface 3
Ethernet
Ethernet
Switch 2
Ethernet
Subnet
1 has
172.30.20.x ThisSwitch
subnet
Subnet
172.30.21.x
one server (Y)
Router
D twoClient
PC (E
R and
Server
and
routers
F) X
172.30.20.
1
C1-…
172.30.20.
47
A1-…
172.30.20.
19
B1-…
Server Y
172.30.21.
86
D1-…
Router E
172.30.21.
2
E1-…
Router F
172.30.21.
1
F1-…
Figure 8-5: Part of an Internet, Continued
Router A
Router B
Arriving
Packet
18
Router C
A packet arrives in Interface 1 of Router B.
The router will forward the packet out a different interface.
Subnet
172.30.20.x
Ethernet
Switch 1
Router D
172.30.20.
1
C1-…
Client PC R
172.30.20.
47
A1-…
Ethernet
Switch 2
Server X
172.30.20.
19
B1-…
Server Y
172.30.21.
86
D1-…
Router E
172.30.21.
2
E1-…
Subnet
172.30.21.x
Router F
172.30.21.
1
F1-…
Figure 8-5: Part of an Internet, Continued
Router A
Router B
Here the packet is
sent out Interface 3,
whichEthernet
connects to
Subnet
Subnet
172.30.21.x
Switch
1
19
Router C
Router B Interface 3
Ethernet
Ethernet
Switch 2
Interface 1
172.30.22.9
21-…
Subnet
172.30.21.x
172.30.20.x
bePC
sent
Router DIt must
Client
R to Server X
Server
Y,
172.30.20.
172.30.20.
172.30.20.
Router
1
47E, or
19
Router
C1-…
A1-…F.
B1-…
Server Y
172.30.21.
86
D1-…
Router E
172.30.21.
2
E1-…
Router F
172.30.21.
1
F1-…
Figure 8-5: Part of an Internet, Continued
Router A
Router B
For a packet going to
Server Y,
Router C
Router B Interface 3
Ethernet
Ethernet
Switch 2
EthernetIP address
The destination
Subnet
Switch 1
172.30.20.x is 172.30.21.86
20
Interface 1
172.30.22.9
21-…
Subnet
172.30.21.x
(Server Y, the destination host)
Router D
Client PC R
Server X
172.30.20.
172.30.20.
The packet172.30.20.
is put in a frame
with
1
47 address D1-…
19
Destination
MAC
C1-…
A1-…
B1-…
(Server Y)
Server Y
172.30.21.
86
D1-…
Router E
172.30.21.
2
E1-…
Router F
172.30.21.
1
F1-…
Figure 8-5: Part of an Internet, Continued
Router A
Router B
For a packet going to
Router E, which will
take responsibility for the packet.
Router C
Router B Interface 3
Ethernet
Ethernet
Switch 2
EthernetIP address
Subnet
The destination
Switch 1
172.30.20.x
is the IP address of
21
Interface 1
172.30.22.9
21-…
Subnet
172.30.21.x
the destination host.
Router D
Client PC R
Server X
172.30.20.
172.30.20.
172.30.20.
The packet is put in a frame with
1
47
19
destination
MAC
address
E1-…
C1-…
A1-…
B1-…
(Router E).
Server Y
172.30.21.
86
D1-…
Router E
172.30.21.
2
E1-…
Router F
172.30.21.
1
F1-…
22
Figure 8-6: Multiprotocol Routing
Unix
Server
Old NetWare
Server
Ethernet
LAN 1
IPX/ SNA
SPX
TCP/IP
TCP/IP
Multiprotocol
Most firms have a mix of internetworking
Ethernet
Router
X
architectures (TCP/IP, IPX/SPX, SNA,
etc.).
LAN 3
Ethernet
Consequently,
are multiprotocol
Internal
LAN 2 most routers
routers that route
the packets
of
Router
Y
multiple architectures.
Site A
Site B
Edge
Router Z
Mainframe
The
Internet
WWW
Server
Figure 8-7: Ethernet Switching Versus IP
Routing
23
Switch 2
Ethernet
Ethernet switching is fast and therefore inexpensive.
Port 5 on Switch 1
Switching
For a destination MAC address,
to Port 3 on Switch 2
there is only one match in the table.
Port 7 on Switch 2
This can be found quickly.
to Port 4 on Switch 3
The frame is sent out the port listed in that row.
Switch 1
A1-44-D5-1F-AA-4C
Switch 1, Port 2
B2-CD-13-5B-E4-65
Switch 1, Port 7
Switching Table Switch 1
Port
Station
2
A1-44-D5-1F-AA-4C
7
B2-CD-13-5B-E4-65
5
C3-2D-55-3B-A9-4F
5
D4-47-55-C4-B6-9F
5
E5-BB-47-21-D3-56
Figure 8-7: Ethernet Switching Versus IP
Routing, Continued
Router B
IP Routing
Router A
IP Routing Table Router A
Interface
Network
1
60.x.x.x
2
128.171.x.x
1
123.x.x.x
2
60.x.x.x
2
123.x.x.x
Interface
1
Network
60.x.x.x
Interface
2
Router C
Router topologies are meshes.
This gives alternative routes.
A destination IP address will
Match multiple rows.
24
Figure 8-7: Ethernet Switching Versus IP
Routing, Continued
Router B
IP Routing
Router A
IP Routing Table Router A
Interface
Network
1
60.x.x.x
2
128.171.x.x
1
123.x.x.x
2
60.x.x.x
2
123.x.x.x
Interface
1
Network
60.x.x.x
Interface
2
Router C
All matching rows must be found.
Then, the best match must be found.
This is slow and therefore expensive.
25
Figure 8-7: Ethernet Switching Versus IP
Routing, Continued

Ethernet (and most other) switching is
inexpensive for a given traffic volume

Router routing is expensive for a given traffic
volume

Network administrators say “Switch where you
can; route where you must.”
26
27
Figure 8.8: Routing Table
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
NextInterface
Hop
Router
1
128.171.0.0
255.255.0.0 (/16)
47
2
G
2
172.30.33.0
255.255.255.0 (/24)
0
1
Local
3
192.168.6.0
255.255.255.0 (/24)
12
2
G
Routers Base Routing Decisions on Their Routing Tables.
Each Row Represents a Route to a Network or Subnet
For Each Arriving Packet,
The Packet’s Destination IP Address
Is Matched Against the
Destination Network or Subnet Field in Every Row
28
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
NextInterface
Hop
Router
1
128.171.0.0
255.255.0.0 (/16)
47
2
G
2
172.30.33.0
255.255.255.0 (/24)
0
1
Local
3
192.168.6.0
255.255.255.0 (/24)
12
2
G
Each Row Represents a Route to a Network or Subnet.
All packets to that network or subnet are governed by that one row.
So there is one rule for a range of IP addresses.
This reduces the number of rows that must be considered.
29
Figure 8.9: Masking
1. Basic Process
Information bit
1 0 1 0
Mask bit
1 1 0 0
Result
1 0 0 0
2. Common Patterns
Binary
Decimal
00000000
0
11111111
255
3. Example 1
IP Address
Mask
Result
4. Example 2
IP Address
Mask
Result
172. 30. 22. 7
255. 0. 0. 0
172. 0. 0. 0
172. 30. 22. 7
255. 255. 0. 0
172. 30. 0. 0
30
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
NextInterface
Hop
Router
1
128.171.0.0
255.255.0.0 (/16)
47
2
G
2
172.30.33.0
255.255.255.0 (/24)
0
1
Local
3
192.168.6.0
255.255.255.0 (/24)
12
2
G
Row 1
If Destination IP Address = 172. 30.33.6
Mask = 255.255. 0.0
Result = 172. 30. 0.0
Destination Network or Subnet = 128.171. 0.0
No match!
31
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
NextInterface
Hop
Router
1
128.171.0.0
255.255.0.0 (/16)
47
2
G
2
172.30.33.0
255.255.255.0 (/24)
0
1
Local
3
192.168.6.0
255.255.255.0 (/24)
12
2
G
Row 2
If Destination IP Address = 172. 30. 33.6
Mask = 255.255.255.0
Result = 172. 30. 33.0
Destination Network or Subnet = 172. 30. 33.0
This row is a match!
32
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
NextInterface
Hop
Router
1
128.171.0.0
255.255.0.0 (/16)
47
2
G
2
172.30.33.0
255.255.255.0 (/24)
0
1
Local
3
192.168.6.0
255.255.255.0 (/24)
12
2
G
Row 3
If Destination IP Address = 172. 30. 33.6
Mask =
Result =
Destination Network or Subnet =
Is this row is a match?
Routing

For Each Incoming IP Packet

Destination IP address is matched against every row
in the routing table.

If the routing table has 10,000 rows, 10,000
comparisons will be made for each packet.

There can be multiple matching rows for a
destination IP address, corresponding to multiple
alternative routes.

After all matches are found, the best match must
be selected.
33
34
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
3
192.168.0.0
255.255.0.0 (/16)
12

NextInterface
Hop
Router
2
If only one row matches, it will be selected as
the best row match.

Destination IP address = 192.168.6.7
G
35
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
13
0.0.0.0
0.0.0.0 (/0)
5

NextInterface
Hop
Router
3
The default row always matches

Mask 0.0.0.0 applied to anything results in 0.0.0.0

This always matches the Network/Subnet value
0.0.0.0

The router specified for this row (H) is the default
router
H
36
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
1
128.171.0.0
255.255.0.0 (/16)
47
2
G
7
128.171.17.0
255.255.255.0 (/24)
55
3
H

NextInterface
Hop
Router
If there are multiple matches, the row with the
longest length of match is selected

This is Row 7 for 128.171.17.56 (24 bit match)

Row 1’s length of match is only 16 bits

Longer matches often are routes to a particular
subnet within a network
37
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
5
172.29.8.0
255.255.255.0 (/24)
34
1
F
8
172.29.8.0
255.255.255.0 (/24)
20
3
H

NextInterface
Hop
Router
If there are multiple rows with the same lengths
of match, the metric column compares
alternative routes.


If the metric is cost, the smallest metric wins (20)
If the metric is speed, the largest metric wins (34)
The Situation

The router first evaluated the IP destination
address of the arriving packet against all rows
and noted the matching rows.

The router then selected the best-match row.

Now, the router examines the interface and
next-hop router fields in the best-match row to
determine what to do with the packet.
38
Figure 8-11: Interface and Next-Hop Router
Router
Forwarding
Packet
Router A
Possible
Next-Hop
Router
Packet to Router B
on Interface 5
Router B
IP Subnet on
Interface (Port 5)
Packet must be sent to
a particular host or
router on the subnet
out a particular
interface (port).
Router C
Possible
Destination
Host
Possible
Next-Hop
Router
39
40
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
5
172.29.8.0
255.255.255.0 (/24)
34

1
F
The Interface specifies the “out” port on the
router.


NextInterface
Hop
Router
A subnet is attached to this interface.
NHR column specifies a specific NHR on that
subnet.

For Row 5, send packet to NHR F on the subnet out
Interface 1.
41
Figure 8.8: Routing Table, Continued
Row
Destination
Network or
Subnet
Mask (/Prefix)
Metric
(Cost)
2
172.30.33.0
255.255.255.0 (/24)
0

NextInterface
Hop
Router
1
Local
If Next-Hop Router Field says Local,

Then the destination host in on the subnet attached to
the interface (1).

Instead of sending the packet to a next-hop router on
the subnet, the router will send the packet to its
destination address.
Routing Recap


The router looks at the destination IP address in
the packet.

First, the router finds all matching rows.

Second, selects the best matching row.

Third, sends packet back out the row’s specified
interface, to the row’s specified next-hop router.
Begins to process the next packet.
42
Quiz

An IP address matches rows 112 and 456.

What row in the routing table will the router look
at first when it searches for matching rows?
(Trick question but one that illustrates a crucial
point.)
43
Quiz

1,000 consecutive packets arrive, all going to
the same destination IP address.

The routing table has 100,000 rows.

This destination IP address matches two rows
in the routing table.

In total, how many rows will the router have to
examine?
44
Routing Recap, Continued

Switches only provide single possible paths, so
there is only one matching entry in the
switching table, and it is quickly found—the one
corresponding to the single path.

Routers have multiple alternative routes and so
must evaluate every row (route) and then select
the best match; this makes routers very
expensive compared to switches for a
comparable traffic volume.
45
46
Figure 8-12: Routing Protocols
Routing
Table
Information
Router
Router
Router
Routers get the information
for their routing tables
by exchanging information
via routing protocols.
Router
Routing
Table
Information
Router
What is “Routing”?

TCP/IP uses the term “routing” in two ways.

First, the forwarding of packets when they
reach a router is called routing.

Second, exchanges between routers in order to
transfer routing table information is called
routing.
47
Figure 8-13: Multiprotocol Label Switching
(MPLS)
Label-Switching Router 1
LabelSwitching
Router 2
Legend
Packet
Label
LabelSwitching
Router 3
LabelSwitching
Router 5
Label-Switching
Multiprotocol Label
Switching (MPLS)
Router
4
can simply forwarding and
therefore
reduce Label-Switched
the cost of router operation.
Path
48
Figure 8-13: Multiprotocol Label Switching
(MPLS), Continued
Label-Switching Router 1
LabelSwitching
Router 2
Legend
Packet
Label
LabelSwitching
Router 3
In multiprotocolLabel-Switching
label switching,
4
a label-switched pathRouter
is determined
for a flow of similar packets.
A label is added before each packet.
LabelSwitching
Router 5
Label-Switched
Path
49
Figure 8-13: Multiprotocol Label Switching
(MPLS), Continued
Label-Switching Router 1
LabelSwitching
Router 2
Legend
LabelLabelSwitching
Packet Label
Label-switching
the way5
Switching routers along Router
look only
Router
3 at a packet’s label,
not at its destination IP address.
Label-Switching
Label-Switching Table
Router
4
Label-Switched
The
label-switching
table tells the router
Label Interface
Path
what interface to use to send
the packet out.
A
1
C
1
F
3
50
Figure 8-13: Multiprotocol Label Switching
(MPLS), Continued
Label-Switching Router 1
LabelSwitching
Router 2
Legend
Packet
Label
Label-Switching Table
Label Interface
A
1
C
1
F
3
Label switching tablesLabelLabelhave only one row per Switching
label.
Switching
Router 5
Router
3 as the row is found,
As soon
the packet can be sent back out.
Label-Switching
Router 4
Label-Switched
As in Ethernet
switching,
Pathinexpensive.
this is fast and therefore
51
Figure 8-13: Multiprotocol Label Switching
(MPLS), Continued
Label-Switching Router 1
LabelSwitching
Router 2
Legend
LabelLabelSwitching
Packet Label
Switching
5
Label switching is similarRouter
to
theRouter
use of 3
virtual circuits in PSDNs.
Label-Switching
Label-Switching Table
Router 4
Label-Switched
Label Interface
Path
A
1
C
1
F
3
52
Figure 8-13: Multiprotocol Label Switching
(MPLS), Continued

MPLS makes transit through an internet much
faster and therefore cheaper than traditional IP
destination address-based routing

In addition, more than one label can be set up
for packets going to a particular network or
subnet

Different labels can give different priorities, etc.

This allows different traffic to be given different
service quality guarantees
53
Figure 8-14: Domain Name System (DNS)
Hierarchy
(root)
Top-Level
Domain Names
.au
.ie 1,
.nlwe saw
.uk that
In Chapter
DNS servers can provide
a target host’s IP address
Second-Level
microsoft.com
hawaii.edu
if you only cnn.com
know its host name.
Domain
Names
However, DNS really is a
Subnetgeneral method for naming
cba.hawaii.edu
Name resources on the Internet.
.edu
.net
.org
voyager.cba.hawaii.edu
.com
Host
Names
ntl.cba.hawaii.edu
54
Figure 8-14: Domain Name System (DNS)
Hierarchy, Continued
(root)
Top-Level
Domain Names
.edu
hawaii.edu
55
.net
.org
.com
.au
.ie
.nl
.uk
Second-Level
microsoft.com
cnn.com
Domain
Names
DNS is organized as a hierarchy.
cba.hawaii.edu
voyager.cba.hawaii.edu
Subnet The top level is the root.
Name
Top-level domains are organized
by type (.com, .edu., etc.)
Host
ntl.cba.hawaii.edu
by
country (.uk, .ie, .ch, etc.)
Names
or by both (.com.us).
http://www.root-servers.org/
Figure 8-14: Domain Name System (DNS)
Hierarchy, Continued
(root)
Top-Level
Domain Names
.edu
.net
.org
Second level domains
.com indicate
.au .ie
.nl
.uk
a company
(cnn.com)
or a product (somemovie.com).
Second-Level
microsoft.com
cnn.com
hawaii.edu
Domain
Companies compete for good
Names
second-level domain names.
(Panko.info, Microsoft.com)
Subnet
cba.hawaii.edu
Name They can get these from
domain name registrars.
voyager.cba.hawaii.edu
Host
Names
ntl.cba.hawaii.edu
56
Figure 8-14: Domain Name System (DNS)
Hierarchy, Continued
(root)
Top-Level
Domain Names
.au .ie
.nl
.uk
At lower levels, more specific
resources can be named.
Second-Level
microsoft.com
cnn.com
hawaii.edu
Domain
One example is the host name.
Names
voyager.cba.hawaii.edu
ntl.cba.hawaii.edu
Subnet cba.hawaii.edu
Name
.edu
.net
.org
voyager.cba.hawaii.edu
.com
Host
Names
ntl.cba.hawaii.edu
57
58
Figure 8-1: Major TCP/IP Standards
User Applications
5 Application
HTTP
4 Transport
3 Internet
SMTP
Many
Others
TCP
Supervisory Applications
DNS
Routing Many
Protocols Others
UDP
IP
ICMP
2 Data Link
None: Use OSI Standards
1 Physical
None: Use OSI Standards
ARP
Note: Shaded protocols are discussed in this chapter.
Figure 8-15: Internet Control Message
Protocol (ICMP) for Supervisory Messages
Router
Host Unreachable
Error Message
IP was created to deliver packets.
Echo Request
(Ping)
ICMP was created to support
Echo
supervisory messages at the internet layer.
Reply
59
Figure 8-15: Internet Control Message Protocol
(ICMP) for Supervisory Messages, Continued
Router
Host Unreachable
Error Message
Echo Request
(Ping)
ICMP messages are
carried in the
data fields
Echo
of IP
packets.
Response
There are no transport
or application layer messages.
ICMP
IP
Message Header
60
Figure 8-15: Internet Control Message Protocol
(ICMP) for Supervisory Messages, Continued
Router
Host Unreachable
Error Message
ICMP error messages
Echo Request
ICMP
IP
advise
senders of delivery
problems.Header
(Ping)
Message
Echo
Reply
This is not reliability;
there is no automatic error correction.
This is only error advisement.
61
Figure 8-15: Internet Control Message Protocol
(ICMP) for Supervisory Messages, Continued
Echo messages can be used toRouter
“ping”
IP addresses or host names.
Host Unreachable
Pinged
hosts
reply with echo reply messages.
Error
Message
This response indicates that the host is active.
Echo (Ping)
Echo
Reply
ICMP
IP
Message Header
62
63
64
Figure 8-16: IPv4 and IPv6 Packets
Bit 0
IP Version 4 Packet
Version Header
Diff-Serv
(4 bits) Length
(8 bits)
Value (4 bits)
is 4
(0100)
Identification (16 bits)
Unique value in each original
IP packet
Time to Live
(8 bits)
Bit 31
Total Length
(16 bits)
Length in octets
Flags Fragment Offset (13 bits)
(3 bits)
Octets from start of
original IP fragment’s
data field
Protocol (8 bits)
Header Checksum
1=ICMP, 6=TCP,
(16 bits)
17=UDP
65
Figure 8-16: IPv4 and IPv6 Packets
Bit 0
IP Version 4 Packet
Bit 31
Version Header
Diff-Serv
Total Length
(4 bits) Length
(8 bits)
(16 bits)
Value (4 bits)
Length in octets
is 4
(0100)
Identification (16 bits)
Flags Fragment Offset (13 bits)
Version (3
field
tells theOctets
version
of start of
Unique value in eachThe
original
bits)
from
the Internet Protocol that the
packetIPfollows.
IP packet
original
fragment’s
data field
The
version
of IPChecksum
today
Time to Live
Protocol
(8dominant
bits)
Header
is Version 4. (IPv4)
(8 bits)
1=ICMP, 6=TCP,
(16 bits)
17=UDP
There were no earlier versions.
66
Figure 8-16: IPv4 and IPv6 Packets, Continued
Bit 0
IP Version
4 Packet
TTL prevents
misaddressed packetsBit 31
Diff-Serv
Totalendlessly.
Length
from circulating
(8 bits)
(16 bits)
Length
in octets
The sender sets
the TTL
value.
Version Header
(4 bits) Length
Value (4 bits)
is 4
(0100)
Each router along the way decrements
TTL value
by 1.
Identification (16 bits) (decreases)
Flags the
Fragment
Offset
(13 bits)
Unique value in each original
(3 bits)
Octets from start of
If a router decrements
to 0,
IP packet
original TTL
IP fragment’s
the router discards the
packet.
data
field
Time to Live
Protocol (8 bits)
Header Checksum
(TTL)
1=ICMP, 6=TCP,
(16 bits)
(8 bits)
17=UDP
67
Figure 8-16: IPv4 and IPv6 Packets, Continued
Bit 0
IP Version 4 Packet
Bit 31
Version Header
Diff-Serv
Total Length
(4
bits)
Length
bits)
bits) data field.
The
Protocol
field tells(8the
receiver what is in the(16
packet’s
Value (4 bits)
Length in octets
1 = an ICMP message
is 4
6 = a TCP segment
(0100)
17 = a UDP datagram
Identification
Offset (13 bits)
There (16
are bits)
other values Flags
for otherFragment
purposes.
Unique value in each original
(3 bits)
Octets from start of
IP packet
original IP fragment’s
data field
Time to Live
Protocol (8 bits)
Header Checksum
(8 bits)
1=ICMP, 6=TCP,
(16 bits)
17=UDP
68
Figure 8-16: IPv4 and IPv6 Packets, Continued
IP Version 4 Packet
Bit 31
Version Header
Diff-Serv
Total Length
(4 bits) Length
Packets may
(8 be
bits)
fragmented (broken into
(16 bits)
multiple packets)
Value (4 bits)
by routers along Length
the way.in octets
is 4
(0100)
Identification (16 bits)
Flags Fragment Offset (13 bits)
Unique value in each original
(3 bits)
Octets from start of
IP packet
original IP fragment’s
data field
Time to Live
Protocol (8 bits)
Header Checksum
(8 bits)
1=ICMP, 6=TCP,
(16 bits)
The receiving host
reassembles the fragmented packet
17=UDP
using information in the Identification, Flags, and Fragment offset fields.
However, fragmentation is rare and typically indicates a hacker attack.
69
Figure 8-16: IPv4 and IPv6 Packets, Continued
Bit 0
IP Version 4 Packet
Bit 31
Source IP Address (32 bits)
Destination IP Address (32 bits)
Options (if any)
Padding
The source and destination IP address fields
are 32 bits long, of course.
Data Field
70
Figure 8-16: IPv4 and IPv6 Packets, Continued
Bit 0
IP Version 4 Packet
Bit 31
Source IP Address (32 bits)
Destination IP Address (32 bits)
Options (if any)
Padding
Data Field
The sender may add Options fields.
if an option does not end at a 32-bit boundary,
padding is added.
Options are rare and usually indicate attacks.
71
Figure 8-16: IPv4 and IPv6 Packets, Continued
Bit 0
IP Version 4 Packet
The data field contains a TCP segment,
UDP datagram,
message,
Source
IP AddressICMP
(32 bits)
or other content.
Destination IP Address (32 bits)
Options (if any)
Data Field
Bit 31
Padding
72
Figure 8-16: IPv4 and IPv6 Packets, Continued
Bit 0
Version
Value
is 6
(0110)
IP Version 6 Packet
Diff-Serv
(8 bits)
Bit 31
Flow Label (20 bits)
Marks a packet as part of a specific flow
The IETF has defined a new version of IP.
This is Internet Protocol Version 6 (IPv6).
Payload Length
Next Header
Hop Limit
(16 bits) The Version field(8value
bits) is 6 (0110).
(8 bits)
Source IP Address (128 bits)
Destination IP Address (128 bits)
Next Header or Payload (Data Field)
73
Figure 8-16: IPv4 and IPv6 Packets, Continued
Bit 0
Version
Value
is 6
(0110)
IP Version 6 Packet
Bit 31
Diff-Serv
Flow Label (20 bits)
(8 bits)
Marks a packet as part of a specific flow
IPv6 has 128-bit source and destination IP addresses.
This allows many more hosts.
Payload This
Length
Next Header
Limit
is important because
some areas Hop
of the
world
(16 bits)
(8 bits)
(8 bits)
are running out
of IP addresses.
Source IP Address (128 bits)
Destination IP Address (128 bits)
Next Header or Payload (Data Field)
74
Figure 8-16: IPv4 and IPv6 Packets, Continued
Bit 0
Version
Value
is 6
(0110)
IP Version 6 Packet
Diff-Serv
(8 bits)
Bit 31
Flow Label (20 bits)
Marks a packet as part of a specific flow
IPv6 adoption has been slow.
IPv4 addressesNext
are not
very scarceHop
yet,Limit
Payload Length
Header
and implementing a (8
new
protocol is difficult
(16 bits)
bits)
(8 bits)
because all routers must be changed.
Source IP Address (128 bits)
However, cellphones, a growing number
of devices
other than
PCs
connected
Destination
IP Address
(128
bits)
to the Internet, and growth in Asia should spur
demand for IPv6 adoption in the future.
Next Header or Payload (Data Field)
75
Figure 8-17: TCP Segment and UDP
Datagram
Bit 0
TCP Segment
Source Port Number (16 bits)
Bit 31
Destination Port Number (16 bits)
Sequence Number (32 bits)
Acknowledgement Number (32 bits)
Header
Length
(4 bits)
Reserved Flag Fields
(6 bits)
(6 bits)
TCP Checksum (16 bits)
Window Size
(16 bits)
Urgent Pointer (16 bits)
Figure 8-17: TCP Segment and UDP
Datagram
Bit 0
TCP Segment
Bit 31
One-bit
fields (16
are bits)
used to characterize
a TCP
segment.
Source
Portflag
Number
Destination Port
Number
(16 bits)
If a bit is “set”, this means that its value is 1.
The flag fields
includeNumber
SYN, ACK,
FIN, and RST.
Sequence
(32 bits)
In order: RST,ACK,PSH,URG,SYN, FIN
010010?
Acknowledgement Number (32 bits)
Header
Length
(4 bits)
Reserved Flag Fields
(6 bits)
(6 bits)
TCP Checksum (16 bits)
Window Size
(16 bits)
Urgent Pointer (16 bits)
76
77
Figure 8-17: TCP Segment and UDP
Datagram, Continued
Bit 0
TCP Segment
Source Port Number (16 bits)
Bit 31
Destination Port Number (16 bits)
Sequence Number (32 bits)
Acknowledgement Number (32 bits)
Header
Length
(4 bits)
The
sequence
Reserved
Flag
Fields number field
Window Size
to be put in (16
order
(6allows
bits) TCP(6segments
bits)
bits)
if IP delivers them out of order
TCP Checksum (16 bits)
Urgent Pointer (16 bits)
78
Figure 8-17: TCP Segment and UDP
Datagram, Continued
Bit 0
TCP Segment
Bit 31
The
Acknowledgement
Number field
tells
the other
Source Port
Number
(16 bits)
Destination
Port
Number
(16 side
bits)
which segment is being acknowledged.
Sequence Number (32 bits)
Acknowledgement Number (32 bits)
Header
Length
(4 bits)
Reserved Flag Fields
(6 bits)
(6 bits)
TCP Checksum (16 bits)
Window Size
(16 bits)
Urgent Pointer (16 bits)
In TCP segments that are acknowledgements,
the ACK bit is set.
79
Figure 8-17: TCP Segment and UDP
Datagram, Continued
Bit 0
TCP Segment
Source Port Number (16 bits)
Bit 31
Destination Port Number (16 bits)
In connection-opening requests,
Sequence Number (32 bits)
the SYN flag bit is set.
Acknowledgement Number (32 bits)
Header
Length
(4 bits)
Reserved Flag Fields
(6 bits)
(6 bits)
TCP Checksum (16 bits)
Window Size
(16 bits)
Urgent Pointer (16 bits)
80
Figure 8-17: TCP Segment and UDP
Datagram, Continued
Bit 0
TCP Segment
Source Port Number (16 bits)
Bit 31
Destination Port Number (16 bits)
Sequence
Number
bits)
In notifications
of(32
closings,
the FIN bit is set.
Acknowledgement Number (32 bits)
Header
Length
(4 bits)
Reserved Flag Fields
(6 bits)
(6 bits)
TCP Checksum (16 bits)
Window Size
(16 bits)
Urgent Pointer (16 bits)
Figure 8-18: Normal Four-Way Closes and
Abrupt Resets in TCP
Normal Four-Way Close
FIN
ACK
FIN
ACK
A normal TCP close is a 4-way close.
81
Figure 8-18: Normal Four-Way Closes and
Abrupt Resets in TCP, Continued
Abrupt Reset
RST
In an abrupt close, one side sends a RST segment
in which the RST bit is set.
The connection is closed by this one segment.
There is no acknowledgements of the RST.
82
83
Figure 8-17: TCP Segment and UDP
Datagram, Continued
Bit 0
TCP Segment
Bit 31
Source Port Number (16 bits)
Destination Port Number (16 bits)
As Module A discusses, the Window Size field
Sequence
can be used in flow
control Number
by telling(32
thebits)
other side
how many more octets it can transmit
before getting
Acknowledgement
another acknowledgement.
Number (32 bits)
Header
Length
(4 bits)
Reserved Flag Fields
(6 bits)
(6 bits)
TCP Checksum (16 bits)
Window Size
(16 bits)
Urgent Pointer (16 bits)
84
Figure 8-17: TCP Segment and UDP
Datagram, Continued
Bit 0
TCP Segment
Bit 31
Source Port Number
(16 bits)transport
Destination
Number (16 bits)
The receiving
processPort
uses
the TCP Checksum field to check the segment for errors.
Sequence Number (32 bits)
If the receiver finds errors, it discards the segment.
Acknowledgement Number (32 bits)
If the segment is correct, the receiver sends an ACK.
Header Reserved Flag Fields
Window Size
Length
(6 bits)
(6 bits)
(16 bits)
(4 bits)
TCP Checksum (16 bits)
Urgent Pointer (16 bits)
85
Figure 8-17: TCP Segment and UDP
Datagram, Continued
In contrast to IP packets,
TCP segments often use options.
TCP Segment
Options (if any)
Data Field
The data field contains an application message, or,
in the case of a supervisory segment, is missing.
Padding
86
Figure 8-17: TCP Segment and UDP
Datagram, Continued
Bit 0
TCP Segment
Source Port Number (16 bits)
Destination Port Number (16 bits)
Sequence Number (32 bits)
Acknowledgement Number (32 bits)
Header
Length
(4 bits)
Bit 31
PortFields
number fields indicate
theSize
Reserved Flag
Window
sending(6
and
receiving application
(6 bits)
bits)
(16processes.
bits)
Similar to the Protocol field in IP packets.
TCP Checksum (16 bits)
Urgent Pointer (16 bits)
Figure 8-19: Use of TCP (and UDP) Port
Numbers

Servers use well-known port numbers for their
major applications.

Port 80 = HTTP

Ports 20, 21 = FTP
 Port 21 for supervisory information
 Port 20 for file transfers

Port 23 = Telnet

Port 25 = SMTP (E-mail)
87
Figure 8-19: Use of TCP (and UDP) Port
Numbers, Continued

Clients Use Ephemeral Port Numbers.


By IETF rules, Ports 49153 to 65535.

Windows follows the rules.

Unix programs usually do not.
The client chooses a random ephemeral port
number for each new connection.
88
Figure 8-19: Use of TCP (and UDP) Port
Numbers, Continued

Registered Port Numbers

Ports 1024 through 49151.

For non-major applications.

Unix does not follow the rules for port number
ranges.
 Unix uses some registered port numbers as
ephemeral port numbers.
89
Figure 8-19: Use of TCP (and UDP) Port
Numbers, Continued

Socket

A socket is an IP address, a colon, and a port
number.
 Example: 128.171.17.13:80

For servers, specifies a specific application on a
specific server.

For clients, specifies a specific connection on a
specific client.
90
Using netstat -n
91
Figure 8-19: Use of TCP (and UDP) Port
Numbers, Continued
Client 60.171.18.22
Ephemeral Source Port Number (50047)
From: 60.171.18.22:50047
To: 1.33.17.13:80
Well-Known Destination
Port Number (80)
A connection has both
a source and destination socket.
Webserver
1.33.17.13
Port 80
Socket is based on the packet IP addresses
and the TCP or UDP port number fieldsSMTP Server
123.30.17.120
Port 25
92
Figure 8-19: Use of TCP (and UDP) Port
Numbers, Continued
Client 60.171.18.22
From: 60.171.18.22:50047
To: 1.33.17.13:80
From: 1.33.17.13:80
To: 60.171.18.22:50047
In two-way communication,
the sockets are reversed
for transmissions in
the opposite direction.
Webserver
1.33.17.13
Port 80
SMTP Server
123.30.17.120
Port 25
93
Figure 8-19: Use of TCP (and UDP) Port
Numbers, Continued
Client 60.171.18.22
From: 60.171.18.22:50047
To: 1.33.17.13:80
If a client connects to two servers,
it will select different ephemeralWebserver
port numbers
(50047 and 60003) for the two1.33.17.13
connections
Port 80
From: 60.171.18.22:60003
To: 123.30.17.120:25
SMTP Server
123.30.17.120
Port 25
94
95
Figure 8-17: TCP Segment and UDP
Datagram, Continued
Bit 0
UDP Datagram
Bit 31
Source Port Number (16 bits)
Destination Port Number (16 bits)
UDP Length (16 bits)
UDP Checksum (16 bits)
Data Field
UDP also uses source and destination port numbers.
The UDP header is very simple because it does
not have to handle connections, error correction,
flow control, and other supervisory matters.
Figure 8-20: Layer 3 Switches and Routers
in Site Internets
To
Other
Sites
Border
Router
Layer 3
Switch
Layer 3 switches are routers.
Layer 3
Switch
Ethernet
However, they are faster than
Workgroup
traditional software-based
Switch
routers because they do
processing in hardware.
Ethernet
Workgroup
Switches
are faster than routers,
Switch
so marketers
invented “Layer 3 switch.
96
Figure 8-20: Layer 3 Switches and Routers
in Site Internets, Continued
To
Other
Sites
Border
Router
Layer 3
Switch
Layer 3 switches are routers.
Layer 3
Switch
Ethernet mean
However, hardware limitations
Workgroup
that they are limited
routers.
Switch
They are not full multiprotocol routers.
Ethernet
They
only support TCP/IP
Workgroup
and,
sometimes, IPX/SPX.
Switch
This limits their usefulness.
97
Figure 8-20: Layer 3 Switches and Routers
in Site Internets, Continued
To
Other
Sites
Border
Router
Layer 3
Switch
Layer 3 switches are routers.
Layer 3
Switch
Ethernet
However, hardware limitations
mean
that they are limitedWorkgroup
routers.
Switch
They usually cannot connect to
Ethernet
WANs because
they usually only implement
Workgroup
Ethernet
at the data link layer.
Switch
A router is normally used at the border.
98
Figure 8-20: Layer 3 Switches and Routers
in Site Internets, Continued
Like traditional routers, L3 switches
require
To considerable management labor.
Router
Other
Therefore, they usually do not
Sites
replace workgroups switches
at the bottom of the hierarchy.
Layer 3
Switch
Layer 3
Switch
Ethernet
Workgroup
Switch
Ethernet
Workgroup
Switch
User
99
Topics Covered

IP

Hierarchical IP addresses

Network, subnet, and host parts

Parts vary in length, but the total is always 32
bits
100
Topics Covered

101
IP


Router Operation

Compare destination IP address of packet to
each row to find all matching rows

Find the best-match row based on length of
match and metric values

Send the packet out the indicated interface to the
indicated destination host or next-hop router
Multiprotocol routers are not limited to routing IP
packets
Topics Covered

IP

Routing Protocols


Allow routers to share route information so they
can update their routing tables
Multiprotocol Label Switching (MPLS)

Bases routing decisions on packet labels instead
of IP addresses

Reduces work compared to normal routing and
therefore costs less
102
Topics Covered


Domain Name System (DNS)

Not just to look up a destination host’s IP address if
you only know its host name

A general system for naming things on the Internet

Firms want second-level domain names (cnn.com)
ICMP

For supervisory messages at the internet layer

Error advisement messages of various types

Pinging to see if a host or router is online
103
Topics Covered


IPv4 Fields

Version

Time to live (TTL)

Protocol

Options (rare and suspicious)

Data field
IPv6

128-bit address fields to allow many more hosts on
the Internet
104
Topics Covered

TCP






One-bit Flag fields (if value is 1, said to be set)
Sequence numbers
Acknowledgement numbers and ACK bit
FIN versus RST closes
Window size field for flow control (Module A)
Port numbers
 Well-known, registered for applications
 Ephemeral for client connections
 Socket syntax = IP address : port number
105
Topics Covered


UDP

Also has source and destination port numbers

Otherwise simple because does not do supervisory
chores
Layer 3 Switches

Routers, but fast and inexpensive like switches.
 But labor cost to manage any router is high

Limited in protocol handling, interfaces

Very attractive where they can be used
106