The implications from this topic are similar in many ways to the implications from last topic. There used to be several distinct protocols used at the network and transport layers but as the Internet has become an important network, most organizations are moving to the adoption of TCP/IP as the single standard protocol at the transport and network layers. This is having many of the same effects described in last topic: the cost of buying and maintaining networking equipment and the cost of training networking staff is steadily decreasing. However, as we move closer to running out of IP addresses, more organizations will move to IPv6. This will cost a lot, but most organizations will see little business value from the change.
As TCP/IP becomes the dominant transport and network layer protocol for digital data, telephone companies who operate large non-TCP/IP-based networks to carry voice traffic are beginning to wonder whether they too should make the switch to TCP/IP. This has significant financial implications for companies that manufacture large networking equipment used in these networks.
SUMMARY
Transport and Network Layer Protocols TCP/IP are the standard transport and network protocols used today. They perform addressing (finding destination addresses), routing (finding the "best" route through the network), and segmenting (breaking large messages into smaller packets for transmission and reassembling them at the destination).
Transport Layer The transport layer (TCP) uses the source and destination port addresses to link the application layer software to the network. TCP is also responsible for segmeting—breaking large messages into smaller segments for transmission and reassembling them at the receiver’s end. When connection-oriented routing is needed, TCP establishes a connection or session from the sender to the receiver. When connectionless routing is needed, TCP is replaced with UDP. Quality of service provides the ability to prioritize packets so that real-time voice packets are transmitted more quickly than simple e-mail messages.
Addressing Computers can have three different addresses: application layer address, network layer address, and data link layer address. Data link layer addresses are usually part of the hardware whereas network layer and application layer addresses are set by software. Network layer and application layer addresses for the Internet are assigned by Internet registrars. Addresses within one organization are usually assigned so that computers in the same LAN or subnet have similar addresses, usually with the same first 3 bytes. Subnet masks are used to indicate whether the first 2 or 3 bytes (or partial bytes) indicate the same subnet. Some networks assign network layer addresses in a configuration file on the client computer whereas others use dynamic addressing in which a DHCP server assigns addresses when a computer first joins the network.
Address Resolution Address resolution is the process of translating an application layer address into a network layer address or translating a network layer address into a data link layer address. On the Internet, network layer resolution is done by sending a special message to a DNS server (also called a name server) that asks for the IP address (e.g., 128.192.98.5) for a given Internet address (e.g., www.kelley.indiana.edu). If a DNS server does not have an entry for the requested Internet address, it will forward the request to another DNS server that it thinks is likely to have the address. That server will either respond or forward the request to another DNS server, and so on, until the address is found or it becomes clear that the address is unknown. Resolving data link layer addresses is done by sending an ARP request in a broadcast message to all computers on the same subnet that asks the computer with the requested IP address to respond with its data link layer address.
Routing Routing is the process of selecting the route or path through the network that a message will travel from the sending computer to the receiving computer. With centralized routing, one computer performs all the routing decisions. With static routing, the routing table is developed by the network manager and remains unchanged until the network manager updates it. With dynamic routing, the goal is to improve network performance by routing messages over the fastest possible route; an initial routing table is developed by the network manager but is continuously updated to reflect changing network conditions, such as message traffic. BGP, RIP, ICMP, EIGRP, and OSPF are examples of dynamic routing protocols.
TCP/IP Example In TCP/IP, it is important to remember that the TCP segments and IP packets are created by the sending computer and never change until the message reaches its final destination. The IP packet contains the original source and ultimate destination address for the packet. The sending computer also creates a data link layer frame (e.g., Ethernet) for each message. This frame contains the data link layer address of the current computer sending the packet and the data link layer address of the next computer in the route through the network. The data link layer frame is removed and replaced with a new frame at each computer at which the message stops as it works its way through the network. Thus, the source and destination data link layer addresses change at each step along the route whereas the IP source and destination addresses never change.
