TCP/IP
Unicast
Unicast addresses are used to send information to a single node on your network. Unicast address types are link-local, unique local and global.
Anycast
Anycast addresses delivers the content to only the closest destination as determined by your router.
Multicast
Muticast addresses send to multiple targets or groups. It is particularly useful supporting broadcast streams.
IPv4
An IP address consists of 32-bits separated in a dotted decimal notation into four octets separated by dots (periods). The numbers are assigned in binary where you would see a number like 11011111.10110010.10011011.00000011. The maximum value is 255 (all 1's in binary) in each octet. The binary number is converted into decimal numbers that are easier for humans to recognize. The example above would translate to a decimal value of 13.168.155.3. Each octet can assign eight binary values of zero or one. A good way to see if your TCP/IP protocol stack has been correctly installed with a new network card you can check the NICs functionality with or without a network connection by pinging it's loopback address. The IPv4 address is 127.0.0.1 and what it does is check the installation and functionality of the interface.
APIPA/link local
Best described as Zero configuration networks APIPA and link local addresses are assigned to hosts when there is no DHCP server reachable and there is no manual address assignment. IPv4 uses the address range of 169.254.1.0 to 169.254.254.255 and IPv6 assigned the block fe80::/64 meaning the address fe80 and enough zeroes to make 64 bits. You may see a % sign followed by a number in the IPv6 link local address. This is a Zone ID network identifier. This is useful as many computers have wired and wireless connections. Each NIC or device on a network segment is given an ID to determine which interface to send particular traffic out on. Link local and APIPA addresses are non-routable.
CIDR
CIDR (Classless Inter-Domain Routing) was established as a more granular way to assign network addresses. This method allowed for the any number of bits to be used as the subnet. Now you could use 27 bits in the mask 255.255.255.224 and end up with eight subnets of 30 hosts each. Now it becomes a matter of tailoring the subnets to your needs. In actual practice you will see CIDR notations on most network addresses. Our last example could be something like 192.168.0.15/27 indicating the 27-bit subnet mask informing you not to assign an address higher than 30. In real life the best way to get this kind of information is by using an online address calculator. These are freely available and essential when planning a network
Gateways and DNS
Network clients can access computers on the local network or internet using its IP address if known. Internet access would be dependent on the configuration of a Gateway on the network. This is a device the handles off network requests and forwards them to the internet. The address of this device is part of the client configuration as is the DNS server address. Now for the average person could know that Googles address is 8.8.8.8, probably not, it is also improbable that they know the IP address of the sites they wish to visit or even devices or machines on the network. Name resolution solves this problem using a DNS (Domain Name Server) to resolve the IP addresses into names. A DNS server maintains a list of the IP addresses of the machines known to it and the domain name. For example www.mysite.com = 202.22.222.2. With that information any user can type in the address by name and the server will look it up and translate the address. If the name is not known to the server it will request it from DNS servers known to it. Once a device successfully connects this information is cached locally. . Remember a Default Gateway and a DNS Server must be configured for your settings to work properly.
Subnet masks
Subnet masks were devised to allow more of the 32-bit address to be used to identify networks. This addressing scheme allows a second set of numbers to identify how many bits of the address to be used to identify the network in eight bit blocks that used to be called classes. The subnet mask is also 32-bits long and formatted the same as the IP address. This allows us to use 16 bits or more for the network address. The 16-bit mask will use all 1's for the first 16-bits for the network address giving you a hexadecimal representation of 11111111.11111111.00000000.00000000 with the zeroes available for host identification. This would be 255.255.0.0 in dotted decimal notation. 24-bit subnet masks are prevalent in SOHO and residential deployment giving you a subnet mask of 255.255.255.0. This was a quick and painless way to provide more networks. The only drawback here is significant as switching from a 24 bit subnet, providing 254 hosts and one subnet, to a 16-bit subnet allows 65534 hosts on one subnet. In most cases this is wasteful and only partially addressed the problem.
Public and Private Addresses
The IANA has reserved the following address ranges for private networks. The best example of private network deployment is the SOHO or simply the residence where you are provided a single public IP address from an internet service provider (ISP) and use NAT/PAT to assign private numbers to use the public address for computers and phones. The address ranges are listed below IPv4- The IANA has reserved the following address ranges for private networks. This means any address within these ranges is private and non-routable. 10.0.0.0 - 10.255.255.255 172.16.0.0 - 172.31.255.255 192.168.0.0 - 192.168.255.255 IPv6- The address block of fc00::/7 is reserved by IANA for private use as Unique Local Addresses(ULA). What this means is that the first block assigned will be fd00::/8.
IP Addresses
The IP protocol is used for host identification. There are two types of IP addresses actively used IPv4 and IPv6. The TCP/IP protocol stack is the foundation of the World Wide Web and most local networks. Each host or client connecting to a network, including the Web, needs a unique address. Duplication of addresses causes a failure of one or both hosts. The IP protocol is used for host identification. There are two types of IP addresses actively used IPv4 and IPv6. The original addressing organization was IPv4 which, while still heavily used today, was recognized to provide an insufficient number of addresses to serve the rapidly growing proliferation of internet connected computers and devices. In 1999 the IANA (Internet Assigned Numbers Authority) began the deployment of the IPv6 protocol.
Client-side DHCP vs Static Addressing
The most common IPv4 or IPv6 TCP/IP addressing method is DHCP (Dynamic Host Configuration Protocol) on a work network there is usually an assigned DHCP server. Residential or SOHO implementations use the router provided by the ISP to perform this function. What DHCP does is assign or lease a unique IP address to each host. The duration of the lease is determined by the network administrator or weekly by default. It will define the internet gateway and Domain Name Server to be used. This means that your machine may not get the same IP address when rebooted or otherwise disconnected. The DNS and Gateway settings will remain. The DHCP server has an available pool of addresses available to assign to clients (hosts) that attempt to connect to the network. The client broadcasts a UDP discovery packet for an address to all connected networks. All DHCP servers will offer an address to the client then the client will accept the offer from the nearest server by requesting a lease and that server will lease that address to the client. The address assignment process is identical for both IPv4 and IPv6 addressing. Only the address formatting will be different. This greatly simplifies network management considering you other option Static IP addressing. Static IP addressing is preferable for Web servers and Internet gateways where you want to be able to consistently locate the machine. Any device that is accessed consistently by users should be statically addressed. Otherwise depending on the size of the network this would be a labor intensive mind numbing task that would require documentation to avoid duplication of addresses. If not memorized you can look up the Gateway and DNS servers on a neighboring machine.
Subnet mask vs. CIDR
The original IPv4 design divided the 32-bit address into four eight bit groups each group had a maximum value of 255. In the original IPv4 design the network number field was allotted the first or most significant eight bits of the address representing the network and the remaining 24 bits (the rest) to identify the host. The problem with this format was while overall it could support over 4 billion devices there were only 254 networks available and was quickly deemed insufficient. A solution was needed to create more networks to support the rapidly growing number of LANs as the internet and business networks were proliferating explosively. A quick solution was needed!
IPv6
To address the inevitable exhaustion of IPv4 addresses the deployment of IPv6 began in 1999 providing 340 undecillion or as a number 340,282,366,920,938,463,463,374,607,431,768,211,456 or plenty in laymen's terms addresses. The IPv6 address is 128-bits in length and although it is actually a binary address the use of a 128-bit binary address is more than impractical for human interpretation. It is expressed in hexadecimal numbers separated by colons (:) no, this is NOT a weird emoji... An IPv6 address will consist of eight groups of hexadecimal values. Hexadecimal numbering allows for up to 16 case sensitive values (0-9 and A-F) to be expressed as opposed to the 10 number decimal system. Each hexadecimal digit represents four digits or bits of binary. Each of the eight groups of an IPv6 address represents 16 bits (Two four bit values). So a typical IPv6 address will look something like this 2002:1F93:AF34:DA4C:0000:0000:0000:CADE. For IPv6 the loopback function is the same as IPv4 using an IPv6 address. This address is represented as 0::1 or even simply ::1. An IPv6 shorthand notation allows for groups of zeroes to be eliminated by using double colons. This would make our example 2002:1F93:AF34:DA4C:: CADE. Note the double colons substituted for the three groups of zeroes. Only one set of colons is allowed in an IPv6 address. Here the binary value of the first block would be 0010 0000 0000 0010. The Interface Identifier or node address is represented by the last 64-bits of the 128-bit address. Our example node is identified as 222 or binary 11011110.