Bitzenbytes.com

IP Addressing

An IP Address is a number that uniquely identifies a TCP/IP host on the Internet or on an Intranet. Host in the parlance of TCP/IP means any machine with a network interface or other communication device configured to use a TCP/IP network. An IP address is a logical address and is not intrinsically tied to any specific hardware until assigned. A computer may have one or more assigned IP addresses depending on the number of installed network interfaces, or the capability of the software that implements the protocol stack. IP addresses were designed to unambiguously specify both the address of a network device and the address of the network to which the device is attached. All IP addresses have two parts: network ID and host ID. However the number of bits used for each address varies according to a complex set of rules for the purpose of allowing flexibility in addressing.

Physical (MAC) Address

The fixed "hardware" address of a computer is the address(s) encoded in firmware on the NIC(s) (Network Interface Card(s)) found in the computer. A computer may have one or more fixed (or physical or hardware) addresses if it has more than one NIC. At the physical layer of a network, any device on the network must communicate using a physical address. A physical address is also called a hardware address or a MAC (media access control) address. Ethernet networks employ a 48 bit physical address.

Terminology: A Media Access Control address is another way of saying any of the following: fixed address, firmware address, hardware address, physical address, MAC address, NIC address.

If a computer uses another method, other than a network adapter with a physical address, to connect to the network, the connection must be made through a communications device that will convert a given communications signal to the signal used on the native media employed by that network's network adapters. In this case, the computer without a physical address can only communicate to the network by virtue of the fact that it communicates through a communications device that does have a physical address.

The IEEE (Institute of Electrical and Electronics Engineers) allocates a 24 bit vendor code to each manufacturer of an Ethernet network interface adapter. The manufacturer is responsible to see that those 3 bytes or octets, which represent the unique identifying code for the manufacturer, is the first 24 bits of every 48 bit address that the manufacturer allocates to any network interface adapter it makes. The second 24 bits is a unique identifier for each device manufactured. This guarantees that no two manufacturers will allocate the same hardware address to a network interface adapter. These 48 bits are written as 6 bytes in hexadecimal notation with dashes between each byte.

When reading this, about IP addressing, if you have any questions about binary, hexidecimal and decimal numbering and their relationships, you should read this.
  

The numbers in physical addresses among several devices on a single network can vary by a lot. In the case of a single network, physical addresses could be ordered and indexed. In the case of more than one network a unique network identifier would have to be attached to the hardware address, and at that point there would be no easy and clear way to logically order the physical addresses that may be found on the communicating networks. Even for a single network, ordered hardware addresses are not easy for people to follow. IP addresses for a single network can be easily ordered. There is no physical reason why one particular IP address must be tied to one particular host. The IP address is a logical address and can be assigned at will either manually as a static address or dynamically.
  

The most important use of an IP address, however, is for routing. Routing is the method by which separate networks can talk. A router is a device that connects two independent LANs. A router is a host (at least dual ported if not multiported--which is to say that a router has more than one IP address, one for a local segment and some number of others to talk to remote segments.) When a client node wishes to connect to a server on a remote network, the connection is made through the router. Quite simply, TCP/IP protocols contain a number of methods used to provide, record, store, retrieve and catalog IP addresses. There is no uniform parallel for physical addresses. The method whereby a host discovers the address of a destination host and the network path that will be followed is called address resolution. Address resolution permits a host address to be discovered without user input. TCP/IP protocols employ IP addresses to determine if information packets should remain only on the originating network or if the information packets should be transmitted to a remote network. There are no common protocols defined that act to allow physical addresses to be employed for this work. TCP/IP is a common protocol employing logical addresses that can be implemented on a wide variety of platforms.

The Format of an IP address

An IP address is an assembly of bits. This address can be represented in binary, hexadecimal, or decimal format, but typically it is represented as dotted decimal notation. An IP address (IPv4) is 32 bits long and can be represented as 4 bytes or 4 octets of 8 bits each. Most often, an IP addresses is assigned or configured in decimal format, but occasionally, it must be recognized in hexadecimal format. This is how an IP address (IPv4) operates at the machine level using binary encoding. In this case of "dotted decimal notation", the four octets are each represented by a decimal number and each number is separated by decimal points. Each octet can have a decimal value between 0 and 255.
  

Table 2:

  
Comparing the table above with the previous table, we can deduct the method employed to discover the number of networks and the number of hosts available for a given IP network address. We'll have a brief look at this now and return to it later.
  

Table 3:

  
Note that no Host ID's can consist of all zeros nor can any Host ID consist of all ones. These addresses are reserved for special addresses. A host ID of all zeros is considered to be the address of a network. A host field with all ones in it is considered to be a broadcast address. Hence, we remove host addresses that consist of all zeros or all ones from the range of available addresses for hosts. There are just two of these for each class and so the formula for discovering the number of hosts available on a network is 2ⁿ - 2, where n is the number of bits available for the Host ID.

A class "A" address is assigned to very large networks with a large number of host computers on the network. It provides 24 bits to define the number of hosts with a Host ID on a given class A network. There are only 126 class A networks available. In this day of Network growth, you can expect that there are none left to be allocated by the InterNIC. Each of the 126 Class A networks can address 16,777,214 (over 16 million) unique host IDs per network. The range of these networks is annotated by substituting letters for possible host IDs. The range of a class A address can be stated as the networks from 1.x.y.z to 126.x.y.z inclusive. In most cases, you can expect that an organization that has been allocated a Class A address will further divide the network using a practice called subnetting. Subnetting allows a single IP address to be segmented to accommodate more than one network segment. Subnetting is applied to all classes for this purpose.

A class B address provides 16 bits to define the number of hosts with a Host ID on a given class B network. A class B address defines 16,384 networks that can each support up to 65,534 hosts per network. The range of a class B address can be stated as the networks from 128.0.y.z to 191.255.y.z inclusive.

A class C address provides 8 bits to define the number of hosts with a Host ID on a given class C network. A class C address defines 2,097,152 networks that can each support up to 254 hosts per network. The range of a class C address can be stated as the networks from 192.0.0.z to 223.255.255.z inclusive.
  

Table 4:

  
Class D addresses are reserved for multicasting. Multicasting is a delivery of a single packet to multiple specified destinations. Only hosts that have been registered to receive the multicast address will accept the packet. Class D addresses have the high order bits of the first octet set to 1110nnnn. Class D addresses cannot be assigned to hosts. Multicasting is meant to be a more efficient method of reaching multiple hosts than Broadcasting. If an IP datagram is broadcast to a subnet, every host on the subnet will receive it, and have to process it, to determine whether the target protocol is active. If it is not, the IP datagram is discarded. Multicasting avoids this overhead by using groups of IP addresses. Each group is represented by a 28-bit number, which is included in a Class D address. The first four bits of a multicast IP address contain the class identification bits. The 28 bits that follow are the multicast address. Multicast group addresses are IP addresses in the range 224.0.0.0 to 239.255.255.255.

For each multicast address there is a set of zero or more hosts which are listening to it. This set is called the host group. There is no requirement for any host to first be a member of a group in order to send to that group. There are two kinds of host group: permanent and transient. For permanent host groups, the IP address is permanently assigned by IANA. The membership of a host group is not permanent; a host may leave or join the group at will. The list of IP addresses assigned to permanent host groups is included in STD 2 - Assigned Internet Numbers (AIN) Some of the important permanent host groups are listed in the Multicast Address Examples below (with the AIN acronym.) A permanent group exists even if it has no members.

Any group which is not permanent is transient and is available for dynamic assignment as needed. Transient groups cease to exist when their membership drops to zero. To join a multicasting host group on a single segment physical network, a process running on a host must somehow inform its network device drivers that it is wishes to be a member of the specified group. The device driver software itself must map the multicast address to a physical multicast address and enable the reception of packets for that address. The device driver must also ensure that the receiving process does not receive any spurious datagrams by checking the destination address in the IP header before passing it to the IP layer.

The basic explanation of IP addresses in this document exposes the rules of IPv4 address space first deployed in 1983. There are two types of IP addresses in active usage: IPv4 and IPv6. IPv6 was first deployed in 1999.  Theoretically, there are 2³² host addresses (4.3 billion) under IPv4, but in practice, the rules for how addresses are formed, and the rules for how addresses are used, reduces the number considerably.   This limitation is recognized and addressed by IPv6.  IPv6 employes 128 bit numbers for IP addresses instead of the 32 bit numbers employed by IPv4. Under IPv6, this works out to about 1.8 x 10²³ addresses for every square foot of land on earth.  This ought to hold us for awhile.  Authority for assignment of IP addresses is delegated.

  
There are three basic ways for a sender to transmit identical data to multiple receivers: broadcast , multiple unicast, and multicast. These methods address the issue of mass distribution of information. We talked about multicasting above. Essentially, Unicast communications are the traditional one message to one destination from one source type; here a source node sends the same message over and over again until all destinations have been reached.

Broadcast communications originate one message from one source and the message is delivered to every host on a network. This is more efficient but it ties up every node in the network with a message whether or not every node needs to get the message.

Multicast messages on a local network segment are less problematic than attempting to multicast messages across routers. Some routers do not support forwarding of multicast packets. Multicast has the benefit of not transmitting packets on network segments that have no designated recipients, yet like a broadcast it transmits a single data stream to multiple receivers.

Any protocol which is connectionless may send broadcast or multicast messages as well as unicast messages. A protocol which is connection-oriented can only use unicast addresses because the connection exists between a specific pair of hosts.
  

Table 5:

Broadcast Addresses

The broadcast address is the address to which devices send packets meant for all other devices. All devices "listen" for broadcasts in addition to their own address. Address Resolution Protocol (ARP) packets and routing information are examples of packets sent to the broadcast address. Most often, the broadcast address is the last address in the network (or subnet), with the host portion being all 1's binary (some networks use 0.0.0.0 or 255.255.255.255, however). Below are some examples of broadcast addresses.
  

Table 6:

  
The first three entries are traditional Class A, B, and C network addresses and use traditional masks. This is usually all that you are shown for broadcast examples and every body expects that the host ID's for broadcasts will be discovered as FFh entries or 255 entries in the IP address. But this is a world where subnetting occurs and the number of bits in a Host ID are not always going to be 8 bits. After you read about subnetting, come back to this explanation. The last two entries in this table are not the usual examples given for broadcast addresses. The fourth entry shows a legitimate Class A broadcast address where the last three octets are not all 1's because of borrowed bits given to a subnetwork. Furthermore, the third octet contains a value that covers both the Subnet ID and the required 1's in the host ID portion of the third octet for the broadcast requirement. Entries 4 and 5 could be "real world"; examples. The fifth entry shows a Class C subnetted address and while there is not 255 entry, nevertheless, the Host ID still is composed entirely of 1's making this a broadcast address. Don't be fooled.

Class E addresses have been reserved for future use and experimental purposes.

Class E addresses have the high order bits of the first octet set to 11110nnn.

Class E addresses cannot be assigned to hosts.

Class addressing is not the only method of IP addressing. We take a brief side trip before returning to class based addressing. In the growing internet world, 32 bit IP addressing is not enough for the expanding number of users with computers who want connectivity.

It appears that there are many more computers to network than there are network addresses. A new network is connected to the Internet every 30 minutes.

Several stopgap measures to avoid running out of addresses have been introduced; new IP addressing schemes have been proposed and are under consideration (IPv6.)

In general, the use of Internet Service Providers is a primary method of distributing the available addresses. ISP's either dynamically allocate IP addresses or assign an IP address to a customer from the block of addresses given to the ISP by the InterNIC. The problem with volume is not only IP addresses for the very large number of computers that want to connect to the Internet. Routing tables are also growing at unprecedented rates. Routing tables are now 30 times the size that they were 10 years ago and they require that the amount of memory devoted to those tables be increased accordingly. 30 times is a very large growth for memory indeed. New Inter-domain Routing paradigms are needed.   One method of reducing the load on routing tables is to use an IP addressing scheme called CIDR (Classless Inter-Domain Routing.) or supernetting. CIDR does not recognize the class of an IP address as indicated by the high order bits of the address; instead it uses a variable length network identifier similar to a subnet mask. CIDR (Classless Inter-Domain Routing) employs hierarchical routing aggregation to minimize routing table entries.

Addressing concerns:

• Plan for the Future: Select a class and subnets that will allow expansion in the future.
• Ensure Uniqueness: When assigning network IDs for an intranet, its important to make sure that each network has a unique network ID. Every segment of your network that is bordered by two router interfaces must have its own unique ID (even if the only two devices on the segment are the two routers.) Each host ID within a network or subnet must be unique.
• Avoid restricted addresses: Certain addresses are not allowed for normal use--don't use them. Host ID's cannot consist of all 0's or all 1's.
• Plan for security: Subnets are one way to restrict privileged traffic.
• Have a method: A logical plan of distribution is useful. Low ID's could be given to Routers. High ID's could be given to servers. The rest might be allocated to workstations. Maybe you should group Unix workstations and Microsoft workstations.
• Mobility: IP addresses must be changed if the host is moved to a new network.
• Record Keeping: is essential.

Internet Addressing:

• A network ID on an internetwork must be unique.
• A network ID cannot be all 0's or all 1's.
• A network ID cannot begin with 127.x.y.z

Private Internets

RFC 1597 - Address Allocation for Private Internets relaxes the rule that IP addresses are globally unique by reserving part of the address space for networks which are used exclusively within a single organization and which do not require IP connectivity to the Internet. There are three ranges of addresses which have been reserved by IANA for this purpose:

1. 10 ... A single Class A network
2. 172.16 through 172.31 ... 16 contiguous Class B networks
3. 192.168.0 through 192.168.255 ... 256 contiguous Class C networks

Any organization may use any addresses in these ranges without reference to any other organization. However, because these addresses are not globally unique, they cannot be referenced by hosts in another organization and they are not defined to any external routers. Routers in networks not using private addresses, particularly those operated by Internet service providers, are expected to quietly discard all routing information regarding these addresses. Routers in an organization using private addresses are expected to limit all references to private addresses to internal links; they should neither advertise routes to private addresses to external routers nor forward IP datagrams containing private addresses to external routers. Hosts having only a private IP address do not have IP-layer connectivity to the Internet. This may be desirable and may even be a reason for using private addressing. All connectivity to external Internet hosts must be provided with application gateways.

Disclaimer: The half life of the information above could possibly be less than that of the average isolated sub-atomic particle.

Please note that the links in the article above were checked within the last two weeks. I note that several have decayed already. Web sites frequently change the location of their documents or remove documents that may not be accessed often. If a link above takes you to a site, but the expected document is not displayed, try searching the site for the document--they may just have moved it.

Alternately, if you take the term that was linked and submit that string to a search engine, it is likely that you will find numerous references to any topic in the post above.