Addressing in TCP/IP
LESSON 6

Outline:
Addressing in TCP/IP

A Need for Identification:

• There is a need to establish:
  • A globally accepted method of identifying computers.
  • To provide a universal communication service.
• TCP/IP universal host identifiers - are standard compact binary addresses used for efficient computation in routing and identification.
  • Scheme is analogous to physical network addressing where:
    • Each host on the internet is assigned an integer address (IP address) of 32-bits.
    • That address is represented in "dotted-decimal" notation.
  • Each address is a conceptual pair of (netid, hostid)
    • netid: identifies a network
    • hostid: identifies a host on that network.

Internet Address Classification:

• IP Addresses (version 4)
  • Five classes (A,B,C,D,E)
  • They are distinguished by their first 5 bits.
  • Addresses are 32-bits.
• Later versions of the Internet Protocol may have more classes and use a different number of bits to represent an address.
  • IPv6 (proposed) uses 64-bits.

• Class E - (11110) - Reserved for future use.
• Class D - (1110) - Multicast addressing.
  • Currently being used with applications that use selective broadcasting in communication:
    • Video-conferencing.
    • MBONE
    • Audio-conferencing.

• The three highest-order bits are used to determine the three Primary Classes:
• Class A - (0) - used for a handful of networks which have more than 65,536 (2^16) hosts attached to each.
  • Netid = 7 bits
  • Hostid = 24 bits
  • Ex. (10.0.0.1) - ARPANET
• Class B - (10) - used for intermediate size networks that have between 256 (2^8) and 65,536 (2^16) hosts attached to each.
  • Netid = 14 bits
  • Hostid = 16 bits
  • Ex. (128.235.251.11) - Hertz.njit.edu
• Class C - (110) smallest configuration for hosts. Allows for less than 256 (2^8) hosts attached to each.
  • Netid = 21 bits
  • Hostid = 8 bits
  • Ex. (195.213.49.3) - Local area networks
• Addresses are designed to allow extraction of either id quickly which makes gateway routing more efficient.
  
  

Network and Host ID Ranges:

Addresses Specify Network Connections:

• Gateways connect two or more physical networks. Each connection requires its own IP address.
• Conventional computers that have two or more physical connections are called multi-homed hosts and require multiple IP addresses.
• IP addresses specify both a network and a host on that network (not an individual machine, but its connection to the network).

Network and Broadcast Addresses:

• Hostid = 0 (128.235.0.0) : names/addresses the network as a whole. (Ex. NJIT network)
• Hostid = All 1's (128.235.255.255) : directed broadcast address which refers to every host on that network. (Ex. Hertz.njit.edu)
• Netid & Hostid = All 1's (255.255.255.255) : limited broadcast address provides broadcast address independent of assigned Netid.
• Directed broadcast addresses require knowledge of the network address, limited broadcast addresses do not.
• Limited broadcast addresses provide a broadcast address for local networks, independent of the assigned Netid.

Network and Broadcast Addresses:
Notations

• 1's => 'All'
• 0's => 'This'
• Netid = 0 => 'This network'

Multicast Addresses:

• Multicast addresses - provide multi-point delivery of packets.
• Also known as selective broadcasting, multicast addressing is a subset of normal broadcasting.
• Multicast addresses can be hardware or software addresses.
• See Lesson 17 (Multicast IP and the MBONE) for further information.

Weaknesses in IP Addresses:

• If a host moves to a different network, it must change it's IP address.
• If the number of hosts grows more than 255 (Class C):
  • The addressing scheme must be changed to a Class B.
  • Extremely time consuming administration changes must also be made to enable the address change.
• Routing:
  • Multiple IP addresses for a single host may be routed entirely different from each other.
  • The route chosen depends upon the address chosen (in some cases).
  • Knowing only one IP address for a single host may sometimes NOT be enough.
• Availability:
  • The number of available IP addresses is rapidly reducing, forcing temporary extensions of the IP addressing scheme, until a more scaleable scheme can be put in place.
  • If a stable replacement for the current IP addressing scheme is not in place within 5-10 years, all IP addresses will be completely exhausted. (IPv6)

Common Representation of Addresses:

• Dotted Decimal Notation - a common way of representing an IP address by using four decimal integers separated by decimal points to ease in reading.
• Each integer represents the value of the binary numbers contained within that octet of the address.
• Example:
  10000000 00001010 00000010 00011110 = 128.10.2.30

Loopback Addresses:

• Class A address 127.0.0.0
  • Reserved for loopback on any local machine to test the communication system and for inter-process communication (IPC) on that machine.
  • Network 127 addresses should never appear on any network, be routed or propagated beyond the network interface card (NIC).
  • Commonly used as local IP addresses when configuring PCs for pseudo-SLIP connectivity.

Obtaining an IP Address:

• Network Information Center (NIC) is the central authority which ensures that all Internet addresses are unique. It does so by managing the approval and assignment of all IP network addresses.
• When requesting an IP address, you should choose a type which fits the current and anticipated future sizes of your company.
• Sample size networks and appropriate address classes:
  • Local Area Networks: Class C
  • Campus-wide Networks: Class B
  • Large Companies: Class B
  • International Networks: Class A
  

• Applying for IP addresses is only needed if your network is going to be connected to the Internet on a permanent basis.
• Currently there is a long waiting list for companies and organizations to obtain IP addresses.
• The NIC may soon charge for IP addresses because of the increased demand and lack of available addresses. Obtaining an IP Address:
• When should an organization apply for an IP address?
  • Any organization that connects to the Internet has no option, they must obtain a registered number.
  • If an organization communicates with it's subsidiaries using the Internet, it too must request a registered number.
• Advantages and Disadvantages of Registering:
  • Advantage:
    • Ensuring that your address and naming conventions are protected (njit.edu).
  • Disadvantage:
    • When you connect to the Internet, your use of a registered address is recorded in the Assigned Numbers RFC along with a contact names of administrators and policy personnel..
• Choosing a network number without official registration:
  • Avoid Class A addresses.
  • Do not copy examples given in a configuration handbook or textbook.
  • Avoid well-known addresses that are published.
  • Use high-end Class B or C addresses at random.

Contacting the Network Information Center:

• Contacting the NIC:

  • Snail Mail:

    Government Systems Inc,

    Attn: Network Information Center

    14200 Park Meadow Drive

    Suite 200

    Chantilly, VA 22021

  • Voice:
    • 1-703-802-4535
    • 1-800-365-DNIC
  • Fax:
    • 1-703-802-8376
  • E-Mail:
    • NIC@NIC.DDN.MIL

Network Byte Ordering:

• Little Endian - lowest memory address contains low- order byte of integer.
• Big Endian - lowest memory address holds the high-order byte of the integer.
• Direct copying may change the value of data from machine to machine.
• TCP/IP protocols define Network Standard Byte Order - most signification byte first (Big Endian)

Mapping Names to IP Addresses:

• Most users do not directly use IP addresses when using network services, they use Domain Names which map to IP addresses.
• When connecting to a server at NJIT, you use the domain name (hertz.njit.edu) and not it's IP address (128.235.251.11).
• Domain names are not IP addresses, but mapped to them for readability by the Domain Naming System (Lesson 13).