Knowledge Base : Glossary : Suggest a Question
Search The Knowledge Base
Browse The Knowledge Base
Knowledge Base : General

TCP/IP Networking Basics

TCP/IP Networking Basics
Chapter 1
About This Manual
Audience, Scope, Conventions, and Formats ................................................................1-1
How to Use this Manual ..................................................................................................1-2
How to Print this Manual .................................................................................................1-2
Chapter 2
TCP/IP Networking Basics
Related Publications .......................................................................................................2-1
Basic Router Concepts ...................................................................................................2-1
What is a Router? ....................................................................................................2-1
Routing Information Protocol ....................................................................................2-2
Internet Protocol (IP) Addresses ....................................................................................2-2
Netmask ...................................................................................................................2-4
Subnet Addressing ...................................................................................................2-4
Private IP Addresses ................................................................................................2-7
Single IP Address Operation Using NAT ........................................................................2-7
Media Access Control (MAC) Addresses and Address Resolution Protocol ...........2-9
Related Documents ..................................................................................................2-9
Domain Name System (DNS) Server .......................................................................2-9
IP Configuration by DHCP ............................................................................................2-10
Internet Security and Firewalls .....................................................................................2-10
What is a Firewall? .................................................................................................2-10
Ethernet Cabling ...........................................................................................................2-12
Category 5 Cable Quality .......................................................................................2-12
Inside Twisted Pair Cables .....................................................................................2-13
Uplink Switches, Crossover Cables, and MDI/MDIX Switching .............................2-14

TCP/IP Networking Basics
A network in your home or small business uses the same type of TCP/IP networking that is used
for the Internet. This manual provides an overview of IP (Internet Protocol) networks and
networking.
Related Publications
As you read this document, you may be directed to RFC documents for further information. An
RFC is a Request For Comment (RFC) published by the Internet Engineering Task Force (IETF),
an open organization that defines the architecture and operation of the Internet. The RFC
documents explain the standard protocols and procedures for the Internet. The documents are
listed on the World Wide Web at http://www.ietf.org and can also be found on many other Web
sites.
Basic Router Concepts
Large amounts of bandwidth can be provided easily and relatively inexpensively in a local area
network (LAN). However, providing high bandwidth between a local network and the Internet can
be very expensive. Because of this expense, Internet access is usually provided by a slower-speed
wide-area network (WAN) link such as a cable or DSL modem. For the WAN link to work on the
Internet, the data traffic meant for the Internet needs to be separated from other WAN data and
forwarded. A router usually performs the tasks of selecting and forwarding this data.
What is a Router?
A router is a device that forwards traffic between networks based on network layer information in
the data and on routing tables maintained by the router. In these routing tables, a router builds up a
logical picture of the overall network by gathering and exchanging information with other routers
in the network. Using this information, the router chooses the best path for forwarding network
traffic.
Routers vary in performance and scale, number of routing protocols supported, and types of
physical WAN connection they support.
TCP/IP Networking Basics
2-2 TCP/IP Networking Basics
v1.0, September 2005
Routing Information Protocol
One of the protocols used by a router to build and maintain a picture of the network is the Routing
Information Protocol (RIP). Using RIP, routers periodically update one another and check for
changes to add to the routing table. RIP-2 supports subnet and multicast protocols. RIP is not
required for most home applications.
Internet Protocol (IP) Addresses
Because TCP/IP networks are interconnected across the world, each computer on the Internet must
have a unique address (called an IP address) to make sure that transmitted data reaches the correct
destination. Blocks of addresses are assigned to organizations by the Internet Assigned Numbers
Authority (IANA). Individual users and small organizations may obtain their addresses either from
the IANA or from an Internet service provider (ISP). You can contact IANA at http://
www.iana.org.
The Internet Protocol (IP) uses a 32-bit address structure. The address is usually written in dot
notation (also called dotted-decimal notation), in which each group of eight bits is written in
decimal form, separated by decimal points.
For example, the following binary address:
11000011 00100010 00001100 00000111
is normally written as:
195.34.12.7
The latter version is easier to remember and easier to enter into your computer.
In addition, the 32 bits of the address are subdivided into two parts. The first part of the address
identifies the network, and the second part identifies the host node or station on the network. The
dividing point may vary depending on the address range and the application.
There are five standard classes of IP addresses. These address classes have different ways of
determining the network and host sections of the address, allowing for different numbers of hosts
on a network. Each address type begins with a unique bit pattern, which is used by the TCP/IP
software to identify the address class. After the address class has been determined, the software
can correctly identify the host section of the address. The figure below shows the three main
address classes, including network and host sections of the address for each address type.
TCP/IP Networking Basics
TCP/IP Networking Basics 2-3
v1.0, September 2005
The five address classes are:
• Class A
Class A addresses can have up to 16,777,214 hosts on a single network. They use an 8-bit
network number and a 24-bit node number. Class A addresses are in this range:
1.x.x.x to 126.x.x.x.
• Class B
Class B addresses can have up to 65,354 hosts on a network. A Class B address uses a 16-bit
network number and a 16-bit node number. Class B addresses are in this range:
128.1.x.x to 191.254.x.x.
• Class C
Class C addresses can have up to 254 hosts on a network. A Class C address uses a 24-bit
network number and an 8-bit node number. Class C addresses are in this range:
192.0.1.x to 223.255.254.x.
• Class D
Class D addresses are used for multicasts (messages sent to many hosts). Class D addresses are
in this range:
224.0.0.0 to 239.255.255.255.
• Class E
Class E addresses are for experimental use.
This addressing structure allows IP addresses to uniquely identify each physical network and each
node on each physical network.
Figure 2-1
7261
Class A
Network Node
Class B
Class C
Network Node
Network Node
TCP/IP Networking Basics
2-4 TCP/IP Networking Basics
v1.0, September 2005
For each unique value of the network portion of the address, the base address of the range (host
address of all zeros) is known as the network address and is not usually assigned to a host. Also,
the top address of the range (host address of all ones) is not assigned, but is used as the broadcast
address for simultaneously sending a packet to all hosts with the same network address.
Netmask
In each of the address classes previously described, the size of the two parts (network address and
host address) is implied by the class. This partitioning scheme can also be expressed by a netmask
associated with the IP address. A netmask is a 32-bit quantity that, when logically combined (using
an AND operator) with an IP address, yields the network address. For instance, the netmasks for
Class A, B, and C addresses are 255.0.0.0, 255.255.0.0, and 255.255.255.0, respectively.
For example, the address 192.168.170.237 is a Class C IP address whose network portion is the
upper 24 bits. When combined (using an AND operator) with the Class C netmask, as shown here,
only the network portion of the address remains:
11000000 10101000 10101010 11101101 (192.168.170.237)
combined with:
11111111 11111111 11111111 00000000 (255.255.255.0)
equals:
11000000 10101000 10101010 00000000 (192.168.170.0)
As a shorter alternative to dotted-decimal notation, the netmask may also be expressed in terms of
the number of ones from the left. This number is appended to the IP address, following a backward
slash (/), as “/n.” In the example, the address could be written as 192.168.170.237/24, indicating
that the netmask is 24 ones followed by 8 zeros.
Subnet Addressing
By looking at the addressing structures, you can see that even with a Class C address, there are a
large number of hosts per network. Such a structure is an inefficient use of addresses if each end of
a routed link requires a different network number. It is unlikely that the smaller office LANs would
have that many devices. You can resolve this problem by using a technique known as subnet
addressing.
TCP/IP Networking Basics
TCP/IP Networking Basics 2-5
v1.0, September 2005
Subnet addressing allows us to split one IP network address into smaller multiple physical
networks known as subnetworks. Some of the node numbers are used as a subnet number instead.
A Class B address gives us 16 bits of node numbers translating to 64,000 nodes. Most
organizations do not use 64,000 nodes, so there are free bits that can be reassigned. Subnet
addressing makes use of those bits that are free, as shown below.
A Class B address can be effectively translated into multiple Class C addresses. For example, the
IP address of 172.16.0.0 is assigned, but node addresses are limited to 255 maximum, allowing
eight extra bits to use as a subnet address. The IP address of 172.16.97.235 would be interpreted as
IP network address 172.16, subnet number 97, and node number 235. In addition to extending
the number of addresses available, subnet addressing provides other benefits. Subnet addressing
allows a network manager to construct an address scheme for the network by using different
subnets for other geographical locations in the network or for other departments in the
organization.
Although the preceding example uses the entire third octet for a subnet address, note that you are
not restricted to octet boundaries in subnetting. To create more network numbers, you need only
shift some bits from the host address to the network address. For instance, to partition a Class C
network number (192.68.135.0) into two, you shift one bit from the host address to the network
address. The new netmask (or subnet mask) is 255.255.255.128. The first subnet has network
number 192.68.135.0 with hosts 192.68.135.1 to 129.68.135.126, and the second subnet has
network number 192.68.135.128 with hosts 192.68.135.129 to 192.68.135.254.
Figure 2-2
Note: The number 192.68.135.127 is not assigned because it is the broadcast address
of the first subnet. The number 192.68.135.128 is not assigned because it is the
network address of the second subnet.
7262
Class B
Network Subnet Node
TCP/IP Networking Basics
2-6 TCP/IP Networking Basics
v1.0, September 2005
The following table lists the additional subnet mask bits in dotted-decimal notation. To use the
table, write down the original class netmask and replace the 0-value octets with the dotted-decimal
value of the additional subnet bits. For example, to partition your Class C network with subnet
mask 255.255.255.0 into 16 subnets (four bits), the new subnet mask becomes 255.255.255.240.
The following table displays several common netmask values in both the dotted-decimal and the
masklength formats.
Configure all hosts on a LAN segment to use the same netmask for the following reasons:
Table 2-1. Netmask Notation Translation Table for One Octet
Number of Bits Dotted-Decimal Value
1 128
2 192
3 224
4 240
5 248
6 252
7 254
8 255
Table 2-2. Netmask Formats
Dotted-Decimal Masklength
255.0.0.0 /8
255.255.0.0 /16
255.255.255.0 /24
255.255.255.128 /25
255.255.255.192 /26
255.255.255.224 /27
255.255.255.240 /28
255.255.255.248 /29
255.255.255.252 /30
255.255.255.254 /31
255.255.255.255 /32
TCP/IP Networking Basics
TCP/IP Networking Basics 2-7
v1.0, September 2005
• So that hosts recognize local IP broadcast packets
When a device broadcasts to its segment neighbors, it uses a destination address of the local
network address with all ones for the host address. In order for this scheme to work, all devices
on the segment must agree on which bits comprise the host address.
• So that a local router or bridge recognizes which addresses are local and which are remote
Private IP Addresses
If your local network is isolated from the Internet (for example, when using Network Address
Translation, NAT, which is described below), you can assign any IP addresses to the hosts without
problems. However, the IANA has reserved the following three blocks of IP addresses specifically
for private networks:
10.0.0.0 - 10.255.255.255
172.16.0.0 - 172.31.255.255
192.168.0.0 - 192.168.255.255
Choose your private network number from this range. Some NETGEAR products have DHCP
servers that are preconfigured to automatically assign private addresses.
Regardless of your particular situation, do not create an arbitrary IP address; always follow the
guidelines explained here. For more information about address assignment, refer to RFC 1597,
Address Allocation for Private Internets, and RFC 1466, Guidelines for Management of IP
Address Space. The Internet Engineering Task Force (IETF) publishes RFCs on its Web site at
http://www.ietf.org.
Single IP Address Operation Using NAT
In the past, if multiple computers on a LAN needed to access the Internet simultaneously, you had
to obtain a range of IP addresses from the ISP. This type of Internet account is more costly than a
single-address account typically used by a single user with a modem, rather than a router.
NETGEAR products use an address-sharing method called Network Address Translation (NAT).
This method allows several networked computers to share an Internet account using only a single
IP address, which may be statically or dynamically assigned by your ISP.
The router does this by translating the internal LAN IP addresses to a single address that is unique
on the Internet. The internal LAN IP addresses can be either private addresses or registered
addresses. For more information about IP address translation, refer to RFC 1631, The IP Network
Address Translator (NAT).
TCP/IP Networking Basics
2-8 TCP/IP Networking Basics
v1.0, September 2005
The following figure illustrates a single IP address operation.
This scheme offers the additional benefit of firewall-like protection because the internal LAN
addresses are not shown to the Internet connection. This filtering can prevent intruders from
probing your system. However, using port forwarding, you can allow one computer (for example,
a Web server) on your local network to be accessible to outside users.
Figure 2-3
7786EA
192.168.0.2
192.168.0.3
192.168.0.4
192.168.0.5
192.168.0.1 172.21.15.105
Private IP addresses
assigned by user
Internet
IP addresses
assigned by ISP
TCP/IP Networking Basics
TCP/IP Networking Basics 2-9
v1.0, September 2005
Media Access Control (MAC) Addresses and Address Resolution
Protocol
An IP address alone cannot be used to deliver data from one LAN device to another. To send data
between LAN devices, you must convert the IP address of the destination device to its MAC
address. Each device on an Ethernet network has a unique MAC address, which is a 48-bit number
assigned to each device by the manufacturer. The technique that associates the IP address with a
MAC address is known as address resolution. Internet Protocol uses the Address Resolution
Protocol (ARP) to resolve MAC addresses.
If a device sends data to another station on the network and the destination MAC address is not yet
recorded, ARP is used. An ARP request is broadcast onto the network. All stations (computers, for
example) on the network receive and read the request. The destination IP address for the chosen
station is included as part of the message so that only the station with this IP address responds to
the ARP request. All other stations discard the request.
Related Documents
The station with the correct IP address responds with its own MAC address directly to the sending
device. The receiving station provides the transmitting station with the required destination MAC
address. The IP address data and MAC address data for each station are held in an ARP table. The
next time data is sent, the address can be obtained from the address information in the table.
For more information about address assignment, refer to the IETF documents RFC 1597, Address
Allocation for Private Internets, and RFC 1466, Guidelines for Management of IP Address Space.
For more information about IP address translation, refer to RFC 1631, The IP Network Address
Translator (NAT).
Domain Name System (DNS) Server
Many of the resources on the Internet can be addressed by simple descriptive names such as
http://www.NETGEAR.com. This addressing is very helpful at the application level, but the
descriptive name must be translated to an IP address in order for a user to actually contact the
resource. Just as a telephone directory maps names to phone numbers, or as an ARP table maps IP
addresses to MAC addresses, a DNS server maps descriptive names of network resources to IP
addresses.
When a computer accesses a resource by its descriptive name, it first contacts a DNS server to
obtain the IP address of the resource. The computer sends the desired message using the IP
address. Many large organizations, such as ISPs, maintain their own DNS servers and allow their
customers to use the servers to look up addresses.
TCP/IP Networking Basics
2-10 TCP/IP Networking Basics
v1.0, September 2005
IP Configuration by DHCP
When an IP-based local area network is installed, each computer must be configured with an
IP address. If the computers need to access the Internet, they should also be configured with a
gateway address and one or more DNS server addresses. As an alternative to manual
configuration, Dynamic Host Configuration Protocol (DHCP) is a method by which each
computer on the network can automatically obtain this configuration information. A device on the
network may act as a DHCP server. The DHCP server stores a list or pool of IP addresses, along
with other information (such as gateway and DNS addresses) that it may assign to the other
devices on the network. Some NETGEAR products can act as DHCP servers.
Some NETGEAR products also function as DHCP clients when connecting to the ISP. Such
NETGEAR products can automatically obtain an IP address, subnet mask, DNS server addresses,
and a gateway address if the ISP provides this information by DHCP.
Internet Security and Firewalls
When your LAN connects to the Internet through a router, an opportunity is created for outsiders
to access or disrupt your network. A NAT router provides some protection because by the very
nature of the process, the network behind the router is shielded from access by outsiders on the
Internet. However, there are methods by which a determined hacker can possibly obtain
information about your network or at the least can disrupt your Internet access. A greater degree of
protection is provided by a firewall router.
What is a Firewall?
A firewall is a device that protects one network from another while allowing communication
between the two. A firewall incorporates the functions of the NAT router, while adding features for
dealing with a hacker intrusion or attack. Several known types of intrusion or attack can be
recognized when they occur. When an incident is detected, the firewall can log details of the
attempt, and it can optionally send e-mail to an administrator to report the incident. Using
information from the log, the administrator can take action with the ISP of the hacker. In some
types of intrusions, the firewall can fend off the hacker by discarding all further packets from the
hacker’s IP address for a period of time.
TCP/IP Networking Basics
TCP/IP Networking Basics 2-11
v1.0, September 2005
Stateful Packet Inspection
Unlike simple Internet sharing routers, a firewall uses a process called stateful packet inspection to
ensure secure firewall filtering to protect your network from attacks and intrusions. Since userlevel
applications such as FTP and Web browsers can create complex patterns of network traffic, it
is necessary for the firewall to analyze groups of network connection states. Using stateful packet
inspection, an incoming packet is intercepted at the network layer and then analyzed for staterelated
information associated with all network connections. A central cache within the firewall
keeps track of the state information associated with all network connections. All traffic passing
through the firewall is analyzed against the state of these connections to determine whether or not
it will be allowed to pass through or be rejected.
Denial of Service Attack
A hacker may be able to prevent your network from operating or communicating by launching a
Denial of Service (DoS) attack. The method used for such an attack can be as simple as merely
flooding your site with more requests than it can handle. A more sophisticated attack may attempt
to exploit some weakness in the operating system used by your router or gateway. Some operating
systems can be disrupted by simply sending a packet with incorrect length information.
TCP/IP Networking Basics
2-12 TCP/IP Networking Basics
v1.0, September 2005
Ethernet Cabling
Although Ethernet networks originally used thick or thin coaxial cable, most installations currently
use unshielded twisted pair (UTP) cabling. The UTP cable contains eight conductors, arranged in
four twisted pairs, and is terminated with an RJ45 type connector. A normal straight-through UTP
Ethernet cable follows the EIA568B standard wiring as described below.
Category 5 Cable Quality
Category 5 distributed cable that meets ANSI/EIA/TIA-568-A building wiring standards can be a
maximum of 328 feet (ft.) or 100 meters (m) in length, divided as follows:
20 ft. (6 m) between the hub and the patch panel (if used)
295 ft. (90 m) from the wiring closet to the wall outlet
10 ft. (3 m) from the wall outlet to the desktop device
The patch panel and other connecting hardware must meet the requirements for 100-Mbps
operation (Category 5). Only 0.5 inch (1.5 cm) of untwist in the wire pair is allowed at any
termination point.
A twisted pair Ethernet network operating at 10 Mbits/second (10BASE-T) will often tolerate lowquality
cables, but at 100 Mbits/second (10BASE-Tx) the cable must be rated as Category 5, or
Cat 5, by the Electronic Industry Association (EIA). This rating will be printed on the cable jacket.
A Category 5 cable will meet specified requirements regarding loss and crosstalk. In addition,
there are restrictions on maximum cable length for both 10- and 100-Mbits/second networks.
Table 2-1. UTP Ethernet cable wiring, straight-through
Pin Wire color Signal
1 Orange/White Transmit (Tx) +
2 Orange Transmit (Tx) -
3 Green/White Receive (Rx) +
4 Blue
5 Blue/White
6 Green Receive (Rx) -
7 Brown/White
8 Brown
TCP/IP Networking Basics
TCP/IP Networking Basics 2-13
v1.0, September 2005
Inside Twisted Pair Cables
For two devices to communicate, the transmitter of each device must be connected to the receiver
of the other device. The crossover function is usually implemented internally as part of the
circuitry in the device. Computers and workstation adapter cards are usually media-dependent
interface ports, called MDI or uplink ports. Most repeaters and switch ports are configured as
media-dependent interfaces with built-in crossover ports, called MDI-X or normal ports. Auto
UplinkTM technology automatically senses which connection, MDI or MDI-X, is needed and
makes the right connection.
The figure below illustrates straight-through twisted pair cable.
The figure below illustrates crossover twisted pair cable.
Figure 2-4
Figure 2-5
TCP/IP Networking Basics
2-14 TCP/IP Networking Basics
v1.0, September 2005
Uplink Switches, Crossover Cables, and MDI/MDIX Switching
In Table 2-1 above on page 2-12, the concept of transmit and receive are from the perspective of
the computer, which is wired as Media Dependant Interface (MDI). In this wiring, the computer
transmits on pins 1 and 2. At the hub, the perspective is reversed, and the hub receives on pins 1
and 2. This wiring is referred to as Media Dependant Interface - Crossover (MDI-X).
When connecting a computer to a computer, or a hub port to another hub port, the transmit pair
must be exchanged with the receive pair. This exchange is done by one of two mechanisms. Most
hubs provide an Uplink switch which will exchange the pairs on one port, allowing that port to be
connected to another hub using a normal Ethernet cable.
The second method is to use a crossover cable, which is a special cable in which the transmit and
receive pairs are exchanged at one of the two cable connectors. Crossover cables are often
unmarked as such, and must be identified by comparing the two connectors. Since the cable
connectors are clear plastic, it is easy to place them side by side and view the order of the wire
colors on each. On a straight-through cable, the color order will be the same on both connectors.
On a crossover cable, the orange and blue pairs will be exchanged from one connector to the other.
Figure 2-6
Note: Flat “silver satin” telephone cable may have the same RJ-45 plug. However, using
telephone cable results in excessive collisions, causing the attached port to be
partitioned or disconnected from the network.
TCP/IP Networking Basics
TCP/IP Networking Basics 2-15
v1.0, September 2005
Some NETGEAR products incorporate Auto Uplink technology (also called MDI/MDIX). With
this feature, each local Ethernet port automatically senses whether the Ethernet cable plugged into
the port should have a normal connection (for example, connecting to a computer) or an uplink
connection (for example, connecting to a router, switch, or hub). That port then configures itself to
the correct configuration. This feature also eliminates the need to worry about crossover cables
because Auto Uplink will accommodate either type of cable to make the right connection.
TCP/IP Networking Basics
2-16 TCP/IP Networking Basics
v1.0, September 2005

Read the attached PDF. for more information with images.

Related Articles
There are no related articles


Viewer Comments Add Comment
No viewers have posted comments

Powered By AutoKB