This is a brief reference suited to the level of Cisco Network Academy. Many Internet data sources give more information. http://en.wikipedia.org/wiki/Internet_Protocol_Suite
hardware
it is all software above Network layer
router >$400,
layer-3 switch >$200
switch $40
patch cord $2, NIC $10, hub (obsolete)
OSI layers
Application
Presentation
Session
Transport
Network
Data Link
Physical
the four layers actually in use on Internet
Application isabove TCP
Application isabove TCP
Application is everything above TCP
TCP
also UDP
IP
also ICMP (ping, traceroute)
ARP
Ethernet & other ways to cable
addresses
protocol data units
segment
packet
frame
bits
like 10 megabits per second or 100Mbps or Gbps
HTTP port 80, SNMP port 161, DHCP port 67, DNS port 53, FTP port 20, TFTP port 69
IP address--for Ethernet port on computer (NIC) or router but not switch ports--IP address is static or dynamic (DHCP)--ver 4 common but ver 6 will be more important soon
MAC address--unique per Ethernet port, from the hardware maker
A concept: if an abbreviation ends in P, it is an abbreviation for protocol, like in IP, TCP, HTTP, SMTP.
A concept: all kinds of packets get sent over the wires of a network and they don't interfere with each other. Various bits in the packet header identify which protocol the packet communicates. All the packets are examined by each NIC, switch, or router, and are ignored by each device that isn't listening for the packet's protocol. If you know what subnetting is, a computer receives all IP packets that are flowing on the subnet and ignores all that have an incorrect network address. A computer that accidentally has the wrong IP address does not disable the subnet, but if there are two hosts with the same IP, one may not work or they both might be intermittent. If your small-office LAN has a DHCP host such as a wireless access point with some wired switch ports, it likely has a range for DHCP addresses that does not cover the whole subnet, and you can do static addressing outside the DHCP range. http://warriorsofthe.net/movie.html is a neat movie that is helpful for people who know at least a little about LANs and the headers used as packets become encapsulated. http://www.firewall.cx/networking-topics/the-osi-model/179-osi-data-encapsulation.html
Meaning of host: you may wonder whether a host means a server. It does not. A host merely has an IP address, which is true for the NIC in a server, client, or workstation. A network printer has an IP address but is not called a host. A router port has an IP address but is generally not called a host. Hub and switch ports don't get IP addresses and are not hosts. Hosts include servers, clients, and computers that do peer to peer.
A concept: if you think that all this networking mumbo-jumbo is way too complex and needs to be scrapped and a fresh, simple start is needed, know that there are large groups of users for every protocol. As complex as it is, it just isn't going to get simpler. That is why IT isn't learned in a month, it is a career. The complex protocols came about in 1973 and were designed by practical people who didn't waste everyone's time with needless complexity.
The complexity of networking is better understood when you have in your mind the schema of the table above.
The private IP address ranges are:
10.0.0.0 through 10.255.255.255
172.16.0.0 through 172.31.255.255
192.168.0.0 through 192.168.255.255