TCP/IP Protocol Suite
Knowing that a protocol is a set of rules and conventions that govern how devices on a network exchange information, this section discusses one of the more commonly used protocol suites, TCP/IP. This discussion will not provide the reader with the amount of information he would need if doing an in-depth study of TCP/IP. Nevertheless, TCP/IP needs to be covered to some degree for the reader to better understand the overall operation of network protocols as the discussions expand in later chapters concerning OSPF.
The TCP/IP protocol suite is also referred to as a stack, and it is one of the most widely implemented internetworking standards in use today. The term TCP/IP literally means Transmission Control Protocol/Internet Protocol. TCP and IP are the two core protocols found within the TCP/IP protocol suite, and their place in the IP protocol stack is clarified in the following paragraphs.
TCP/IP was originally developed for ARPANET, a U.S. Government packet switched WAN, over twenty-five years ago. Although at the time this was a private network and a protocol was designed specifically for use within it, TCP/IP has since grown in popularity and is one of the most open protocols available for use in networks today. This growth and popularity is primarily due to TCP/IPs capability to connect different networks regardless of their physical environments. This has made TCP/IP todays de facto standard.
Because TCP/IP was developed long before the OSI model, it is not 100 percent compatible with the OSI model. However, TCP/IP can run over OSI-compliant lower layers, such as the Data Link and Physical layers of the OSI model. TCP/IP can communicate at the Network layer as well using IP.
TCP is the main transport layer protocol that offers connection-oriented transport services. TCP accepts messages from upper-layer protocols and provides the messages with an acknowledged reliable connection-oriented transport service to the TCP layer of a remote device. TCP provides five important functions within the TCP/IP protocol suite:
IP is the main network layer protocol that offers unreliable connectionless service because it depends on TCP to detect and recover from lost packets. IP provides three important functions within the TCP/IP protocol suite:
The term datagram describes a unit of data at the TCP layer. At the IP layer, it is called a packet, and at the lower layers, it is called a frame.
If a message is too large for the underlying network, it is up to the IP layer to fragment the datagram into smaller parts.
Different paths might be available through the Internet, between a source and a destination station. Fragments of a datagram might take different paths through a network.
So, when messages arrive at the destination station, the IP protocol stack must sequence them and reassemble them into their original datagram.
Each datagram or fragment is given an IP header and is transmitted as a frame by the lower layers.
Common TCP/IP Routing Protocol Characteristics
This section will briefly describe the two most common and relevantto this bookrouting protocols that TCP/IP has available for useRIP and OSPF. A more thorough discussion of these routing protocols will be provided in Chapter 3, Understanding & Selecting Networking Protocols, and throughout other parts of this book.
RIP (Routing Information Protocol) Characteristics
RIP is a distance vector protocol designed at Berkeley in the late 1960s that is still widely deployed today, in which the router only exchanges routing information from the connected neighbors.
OSPF (Open Shortest Path First) Characteristics
OSPF is a link state protocol in which all routers in the routing domain exchange information and thus know about the complete topology of the network. Because each router knows the complete topology of the network, the use of the shortest path first algorithm creates an extremely fast convergence.
Basic Protocol Operations
A routed protocol such as IP is used as the method of communication between devices on a network. Using a selected routing protocol, such as OSPF, which is supported by the routed protocol, such as IP, you can build the network so that every device can communicate. For example, as mentioned previously, IP is a routed protocol that can use either OSPF or RIP as its routing protocol.
The routing protocol used builds routing tables that tell the router the optimal path to a destination. Routers compare metrics (measurements based upon path characteristics) to determine the optimal path, which is then stored in a routing table. Each routing protocol has an algorithm that it uses as a basis for its calculations, which might lead you to ask: Why is an algorithm needed?