IP Address

Reference: [Book] Network Introduction for IT Engineers

• The layers in the OSI 7 layers that have addresses are Layer 2 and Layer 3

  • Layer 2 uses the physical address, the MAC address,

  • Layer 3 uses the logical address, the IP address

• Since most networks operate on TCP/IP, understanding the IP address system is important

• The Layer 3 address of the IP address and other protocol stacks has the following characteristics

  1. It is a logical address that users can change

  2. The address has levels

     • Divided into a network address representing the Group and a host address

1. IP Address System

• The IP addresses we commonly use are 32-bit IPv4 addresses

  • IP has two systems in use: v4 and v6, with IPv6 addresses being 128 bits

• When expressing IPv4 addresses, they are divided into 4 8-bit units called octets, and each octet is separated by "."

  • IPv4 address: 8 bits x 4 = 32 bits

• Unlike Layer 2 MAC addresses expressed in hexadecimal, IP addresses are expressed in decimal, so an 8-bit octet can have values from 0 to 255

• While Layer 2's MAC address is divided into the manufacturer code OUI and manufacturer-specific serial number UAA for different purposes,

  • Layer 3's IP address is also divided into a network address and a host address

    • Network address

      • An address that refers to the network that groups hosts

      • Networks with the same Network address are called Local Network

    • Host address

      • An address to distinguish hosts within a single network

• MAC addresses are split evenly at 24 bits each, but the boundary between the IP address's network address and host address is not fixed

  • This is the most distinguishing feature of the IP address system from other address systems

  • The IP address system introduced the Class concept to allocate different network sizes based on the required number of host IPs

• If the separator dividing network and host addresses is fixed, all networks would be the same size since they could have the same number of host IP addresses,

  • But since the separator can be moved, network sizes can vary

• The class concept introduced by IP addresses has the advantage of being able to conserve addresses compared to other fixed network address systems

  • If all networks were the same size,

    • Organizations needing large networks would face difficulty in securing multiple networks and receiving contiguous networks

    • Organizations needing small networks would take too many IPs, resulting in IP waste

• A, B, and C Classes can be distinguished just by looking at the first octet's address

Classful Network

A Class

• Has the largest address space

• Can have approximately 16 million IP addresses

• The separator between network and host addresses is at the first octet

  • This separator is called the subnet mask!

• Since there is 1 octet for the network address and 3 octets for the host address,

  • It has 2^8 (256) networks and

  • 2^24 (16,666,216) host addresses per network

• A, B, and C classes can be distinguished just by looking at the first octet address; if the octet address is in the range 0 ~ 127, it is A Class

  • In binary, addresses where the first octet is 0 0000000 ~ 0 1111111 are A Class

  • Only 127 is an exception, used as the Loopback address representing itself, so the addresses actually usable as A Class are 1.0.0.0 ~ 126.255.255.255

B Class

• Has approximately 65,000 IP addresses

• The separator is at the second octet

• Since there are 2 octets for the network address and 2 octets for the host address,

  • It has 2^16 (65,536) networks and

  • 2^16 (65,536) host addresses per network

• Addresses where the first digit is 1 and the second digit is 0 in binary representation of the first octet are B class

  • Binary first octet is 10 000000 ~ 10 111111,

  • In decimal, addresses from 128 to 191 are B Class

C Class

• The separator is at the third octet

• Since there are 3 octets for the network address and 1 octet for the host address,

  • It has 2^24 (16,777,216) networks and

  • 2^8 (256) hosts per network

• Addresses where the first and second digits are 1 and the third digit is 0 in binary representation of the first octet are C Class

  • Binary first octet is 110 00000 ~ 110 11111,

  • In decimal, IPs from 192 to 223 are C class

The class-based network division technique is a concept used in the past, and networks are no longer divided based on classes

To more precisely allocate network addresses, a method of detailed division in 1-bit units is used to match the required network size

+

Determining the Number of Usable Hosts in a Network

• In IP networks, since network sizes change, it is important to determine the number of hosts and valid IP range usable in a single network

  • When networks are divided by Class, they are expressed in easily recognizable decimal form making it straightforward to understand, but for Classless Networks, determining the valid IP range is very important

• Generally, the total count of all expressible numbers can be calculated as base^digits

  • ex) The total numbers a 4-digit decimal number can represent is 10^4

    • 10,000 numbers from 0000 to 9999

• Since IP uses binary, the number of expressible IPs can be calculated as 2^digits

  • One octet is 8 binary digits, so A Class has 2^24, B Class has 2^16, and C Class has 2^8 IPs

  • However, in the IP system, the first number is used as the network address and the last number as the broadcast address, so the actually usable IPs are A Class 2^24 - 2, B Class 2^16 - 2, C Class 2^8 - 2

2. Classful and Classless Network

• When the IP address system was first created, introducing the Class concept was a good choice that could provide scalability and minimal address waste

  • In Classful Networks, there was no need for a separator (Subnet mask) to divide network addresses and host addresses

  • Simply looking at the first digit naturally told which class the address belonged to, and the address separator could be applied

Classless Network

• As the Internet became commercialized, the number of hosts connected to the Internet increased explosively

  • The existing Classful address system was a good system capturing both scalability and efficiency, but it was far too insufficient to handle the exponentially growing IP address demand

    • Additionally, many IPs were wasted in the process of hierarchizing and dividing network addresses

• The 3 conservation and transition strategies created to solve IP address shortage and waste problems are as follows

  1. Short-term measure - Classless, CIDR (Classless Inter-Domain Routing) based address system

  2. Medium-term measure - NAT and Private IP addresses

  3. Long-term measure - Next-generation IP IPv6

• The biggest problem with IPv4 was not just the shortage of addresses themselves, but the waste from organizations that received upper Classes (A Class) not properly using those addresses

  • In Classful, once a single class network was assigned to an organization, no matter how many addresses were unused, IPs could not be divided for other organizations to use

    • To solve this problem, the class concept itself was abandoned, and this is called Classless network

      • The address system we currently use is a Classless-based address system that does not apply the Class concept

• In Classless Network, a separate separator must be used to divide network and host addresses, and this separator is called the Subnet Mask

  • The Subnet Mask is used to distinguish between IP addresses and Network addresses, where binary 1 represents Network address and 0 represents Host address

    • Typically expressed in decimal as 255.0.0.0, 255.255.0.0, 255.255.255.0

    • Since binary 11111111 in decimal is 255, 255 represents the Network address portion and 0 represents the Host address portion

      • ex) An IP of 103.9.32.146 with subnet mask 255.255.255.0 has

        • Network address of 103.9.32.0,

        • Host address of 0.0.0.146

  • In Classless-based IT Networks, the subnet mask is required to represent the Network, and it must be used when assigning IP addresses to servers or PCs

Subnet Mask Notation Methods

• There are two methods for expressing Subnet masks: bit-level notation and decimal notation

  • Bit-level notation expresses the number of consecutive 1s in the Subnet Mask

    • Expressing A Class as a subnet mask: the first Octet is 1 and the remaining Octets are 0, so it is expressed as /8

    • B Class is /16,

    • C Class is expressed as /24

  • Decimal notation:

    • A Class is 255.0.0.0,

    • B Class is 255.255.0.0,

    • C Class is 255.255.255.0

3. Subnetting

• Ignoring the originally assigned Class criteria and having the user define new Network-Host criteria to subdivide the network smaller than classful units is called Subnetting

  • It is called subnetting because the assigned address is re-divided, and this is the most significant feature of classless networks

  • Subnetting divided by Octet units is easy to understand and operate, but in practice, networks are divided more finely than Octet units, in 1-bit units of binary

• In practice, there are two situations where subnetting must be considered, and the factors and scope to consider differ depending on the situation

  1. When a network designer plans how to efficiently divide the network

     • Network scope design considering the required terminals within the network

  2. When a network user needs to distinguish between their own network and remote networks in an already divided network

     • Determining the usable IP range in the network

     • Verifying that the Default Gateway and Subnet Mask settings are correct

3-1. Network User's Subnetting

• Network users need to determine the usable IP address range in an already designed network

  • If an IP outside the given network range is assigned, or the Subnet mask is incorrectly entered,

    • Communication problems may occur with terminals in a specific range of the Local Network,

    • Or communication with the entire external Network may fail

• Most subnetting is done in bit units, and since the IP address system is processed by computers, it is in binary

  • If you are familiar with binary or subnetting is done in octet units, it is easy to determine the network size and IP range, but for 1-bit subnetting, determining the valid network range is difficult

    • Generally, the method to determine the valid range of your own Network is as follows

      1. Express your IP in binary

      2. Express the Subnet Mask in binary

      3. Use a binary AND operation to find the subnetted Network address

      4. Change all Host address bits to binary 1 to find the Broadcast address

      5. Determine the valid IP range

         • Subnetted Network address + 1 is the smallest valid IP

      6. Broadcast address - 1 is the largest valid IP

      7. Convert the binary results to decimal

    • It is not always possible to determine the Local Network range using the above method

Simple Subnetting Method

• ex) IP address: 103.9.32.146, Subnet Mask: 255.255.255.192

  1. Convert the Subnet Mask to binary

     • ex) 11111111.11111111.11111111.11000000

  2. Determine the maximum number of IPs the current Subnet can have

     • ex) 2^6 = 64

  3. List multiples of 64 to identify the base Network addresses

     • The first Block starts from 0

     • The last address of each Network becomes the Broadcast address

       • This address is the next block's Network address - 1

     • ex) 0 ~ 63 / 64 ~ 127 / 128 ~ 191 / 192 ~ 255

  4. Select the Network that contains Host address 146 from 103.9.32.146

     • ex) 128 ~ 191

  5. Organize the required addresses

     • Network address: 103.9.32.128 (first number)

     • Broadcast address: 103.9.32.191 (last number)

     • Valid IP range: 103.9.32.129 ~ 103.9.32.190 (between Network and Broadcast addresses)

The key to the above method is determining the Network size based on the Subnet Mask to find the subnetted Network size!

3-2. Network Designer's Subnetting

• When building a new Network, contrary to the network user, the designer is not given a Subnet Mask but must consider the Network size to determine the Subnet Mask and reflect it in the design

  • What a network designer must consider when designing IPs:

    1. How many IPs need to be allocated in a single subnetted Network?

    2. How many subnetted Networks are needed?

ex) Designing a Network for a company with 12 branches

1. 12 IPs must be allocated in a single subnetted Network

2. Networks grow in powers of 2, so they can be allocated in units of 4, 8, 16, 32, 64, 128, 256

   • The smallest Network that can accommodate 12 IPs is 16, so allocate a Network of 16

     • However, a Network of 16 must exclude 2 IPs for Network address and Broadcast address, so the actually usable IPs are 14

3. Secure 12 of the 16-unit Networks

   • List multiples of 16 starting from 0 to identify Network addresses

4. Allocate 12 Networks out of the total 16 to each branch

• When designing a Network, it is best to use private IP ranges to ensure sufficient IP range

• If public IPs are used with no margin for Network allocation, there will be many different-sized Networks

  • This makes management difficult for network administrators and makes it hard for general users to easily distinguish or recognize IPs

• It is advisable to allocate same-sized Networks as much as possible, and divide in C class units of 24 bits that are easily understood even in decimal notation

4. Public IP and Private IP

• To access the Internet, an IP address is needed and this IP must be a globally unique identifier

  • Such an IP address is called a Public IP address

• If you configure a Network privately without connecting to the Internet, you can build a Network without being assigned a public IP address

  • The IP addresses used in this case are called Private IP addresses

• To access the Internet, you need to be assigned an IP address from a telecommunications carrier or obtain an Autonomous System Number (ASN) from an IP allocation agency (KISA in Korea) and then receive independent IPs, making the process complex

• If you do not access the Internet or use NAT (Network Address Translation) technology (using a home router or company firewall), private IPs can be used

  • These addresses are specified in the Internet standard document RFC

  • Using Private IP prevents direct Internet access, but after converting to public IP at the NAT device, Internet access becomes possible

• When building a Private Network within a company, even if connecting to the Internet via NAT, IPs assigned to other users should not be used as private IPs

  • If Public IPs used by other organizations are used internally, access to those IPs becomes impossible

• Private IPs can use 1 A Class, 16 B Class, and 256 C Class

  • Large Enterprise networks mostly use the A Class size 10.0.0.0/8 network,

  • For smaller networks, C Class 192.168.x.0/24 is used

    • This is why the most common default IP for home routers is 192.168.0.1