TCP/IP BASICS
- Intro to IP Addressing & Binary
- Each computer on a TCP/IP network must have a unique IP address.
- IPv4 addresses are written as four octets, such as: 192.168.4.12
- Each octet represents a binary string;
- 192, for example, is represented by 11000000
- Dotted decimal notation–shorthand used to represent the 32 1’s & 0’s.
- 28 (256) combinations
- Each octet is valued between 0 and 255.
- Converting from binary to dotted decimal:
- Intro to ARP
- APR (Address Resolution Protocol) resolves IP addresses.
- ARP is what a computer uses when it knows the IP address, but needs the MAC address.
- Type ‘arp -a’ to see the ARP cache (shows a list of the IP addresses & the MAC addresses).
- ARP requests are broadcast over a network
- Subnet Masks
- Each host needs a subnet mask.
- The host uses the subnet mask to know if the destination is on the local network or a remote network.
- Each host knows the default gateway so that it can forward traffic to remote networks. The default gateway will figure out where to forward the message.
- Network ID–the part of the network numbering system that has to be identical for every computer/device on that network.
- Host ID–the part that changes for each individual computer/device.
- Cannot use 0 or 255 for the Host ID; can use any number from 1 to 254.
- Note: Subnets CAN end in 0 or 255.
- Note: The subnet mask is only used by the computer–it is never sent out.
- Note: Usually the subnet mask gets smaller for really huge networks & longer (like a /24 [“whack 24”] for smaller networks).
- *Remember: When working with computers the most important thing to remember is:
- You’re going to have to have an IP address.
- You’re going to have to have a subnet mask.
- You’re going to have to have a default gateway (aka your router!!!)
- Classful Addressing
- Remember, the most important thing to remember about the internet is that every single computer on the internet HAS to have a real IP address!
- Classful subnetting was the first effor to divide network ID’s.
- Class A, B, & C licenses.
- Memorize the first octet to know your class licenses!
- The IANA (Internet Assigned Numbers Authority) keeps track of all the IP address & delegates them as needed.
- The IANA passes numbers/(IP addresses) to RIR
- RIR—Regional Internet Registry–worldwide network broken up geographically
- AfriNIC (African Network Information Center)
- APNIC (Asia Pacific NIC)
- ARIN (American Registry for Internet Numbers)
- LACNIC (Latin American & Caribbean NIC)
- RIPE NCC (Reseaux IP Europeans Network Coordination Centre)
- The RIRs then pass out chunks of IP addresses to ISPs. (Internet Service Providers)
- ISP’s then pass off IP address to consumers.
- “Class licenses” were used to organize all of this.
- Class A 0-126 /8 (Any number from 0 to 126, “wack 8”)
- ex: 14.0.0.0 (can be used to create about 16.7 million different combinations for IP addresses!) (Comcast or other big customers.)
- Class B 128-191 /16 (Class B starts with 128 up to 191 and are “wack 16” addresses.)
- Class B creates about 65,534 IP addresses.
- Class C 192-223 /24 (Class C always starts with a 192 up to a 223, and are always “wack 24’s”. ex: 193.44.16 /24
- So we can always identify a Class A, B, or C simply by memorizing the first number in the octet!
- Subnetting divides Network IDs into two or more networks.
- Subnets don’t have to be on the “dots” of the dot notation.
- Subnetting With CIDR
- CIDR (Classless Inter-Domain Routing) pronounced “cider”.
- Subnet masks have all 1’s on the left and all 0’s on the right.
- Remember, the dotted decimal notation system is just for humans to better read the info. The computers just see a long string of 1’s & 0’s.
- The more subnets you have, the less hosts are available.
- CIDR has been around since the 90’s, it’s a powerful & complicated tool.
- CIDR gives more granular control over how we take one subnet & chop it up into lots of little subnets.
- We, as individual consumers, don’t do a lot of subnetting ourselves, BUT we enjoy the benefits of it. Subnetting is most commonly done by ISPs.
- REMEMBER, subnets don’t have to be all 255s & 0s.
- Subnets have to be a string of ones (1’s) followed by a string of 0’s!
- Also, remember, that the more you subnet something, the less hosts you have.
- 2 subnets, 126 host per sub
- Dynamic vs Static IP addresses
- /24 (“wack 24”) = 254 hosts
- Subnet Mask- /24 = 254 possible IP addresses
- Knowing the number of hosts you need helps determine how small of a CIDR subnet would be needed.
- /25 = 126 hosts (or 27 = 128-2 = 126 (We subtract 2 because you can’t use all 0’s or all 1’s as a real address, remember!)
- /26 = 62 hosts (26 = 64-2 = 62)
- /27 = 27-1’s followed by 5-0’s; so 25 = 32-2 = 30 hosts
- /28 = 14 host (24 =16-2 = 14)
- /29 = 6 hosts (23 = 8-2 = 6)
- /30 = 2 hosts (22 = 4-2 = 2)
- /31 = 0 hosts (21 = 2-2 = 0)
- It’s really important that if we remember our subnet sizes, we can pretty much instantaneously know the number of possible hosts!
- The ISP should provide the numbers of the subnet mask & Network ID.
- Dynamic & Static IP Addressing
- Each broadcast domain must have only one DHCP server.
- Every modern operating system comes with DHCP enabled by default.
- DHCP Relay enables a single DHCP server to service more than one broadcast domain.
- DHCP & BOOTP (Bootstrap Protocol)–Linux
- Process ex:
- DHCP Discover broadcast message (Remember, broadcast is a MAC address of all F’s!) sent by DHCP client (computer).
- Looking for DHCP server somewhere on the network.
- The DHCP server will receive the ‘Discover’ broadcast message & send back a DHCP Offer (as a unicast message back to the client directly).
- all the info needed (the IP address; the subnet mask; the default gateway; and other info)
- The requesting client computer will receive the offer and then send a DHCP Request back to the DHCP server. (Basically saying ok, I’ll use that info!)
- Once the DHCP server gets the request, he sends a ‘DHCP Acknowledge‘ back to confirm and stores all this info & keeps track of along with all the other clients on the network that are using DHCP.
- The important part to remember is that for all this to happen, a DHCP server is needed! (Almost all home routers now include this already built-in.)
- Notes:
- Each broadcast domain must have only one DHCP server.
- We DON’T want 2 servers responding to requests with different info!!!)
- DHCP server has to be run within broadcast domain. (It can’t be located outside of the network, (unless using a DHCP relay!).)
- Rogue DHCP Servers
- If you get an APIPA (Automatic Private IP Addressing) address, check to see if you are connected to a DHCP server.
- If you are connected to a DHCP server, and still get an APIPA address, make sure the DHCP server is working.
- If you get an IP address other than your correct network ID, you may have a rogue DHCP server.
- All DHCP problems manifest the same way; you open a browser & can’t access the internet, because you don’t have a valid IP address, or default gateway or DNS.
- APIPA–built into all DHCP clients; designed as a fallback if you can’t find a DHCP server.
- APIPA addresses always start with 169.254.
- Special IP Addresses
- Special internal IP addresses are: ’10.x.x.x’ , ‘172.16.x.x’ – ‘172.31.x.x’ & ‘192.168.x.x’.
- The loopback address for IPv4 is ‘127.0.0.1’ and the loopback address for IPv6 is ‘::1’.
- *An APIPA address (‘169.254.x.x’) indicates the DHCP server is down!!!
- IP Addressing Scenarios
- ‘ipconfig’ (Windows) and ‘ifconfig’ (Linux) display the IP address information.
- Virtual machines can be a source of duplicate MAC address errors.
- All the computers in one broadcast domain have the same subnet mask.
