This article constitutes section 1, part C, chapter 4 of the CIW Website design manager course and briefly covers: TCP/IP Internet Architecture
TCP/IP Internet Architecture
Here we shall look at how TCP/IP is used for the Internet. As with other protocols, the Internet Architecure is split into layers, with each layer corresponding with layers in the OSI/RM.
- Network Access Layer
- The Network Access Layer of the Internet Architecture corresponds to the physical and data link layers of the OSI model. The layer usually consists of: The OS device driver, the interface card and the physical connections.
- Internet Layer
- This layer corresponds to the network layer of the OSI model. Protocols used at the Internet Layer are: Internet Protocol (IP), Internet Control Message Protocol (ICMP), Internet Group Management Protocol (IGMP), Address Resolution Protocol (ARP) and the Reverse Address Resolution Protocol (RARP).
- Transport Layer
- The Transport layer corresponds to the transport and session layers of the OSI model. The transport layer accepts data from the application layer, splits it into packets and provides the flow of information between two hosts. Protocol used at this layer are: Transmission Control Protocol (TCP) and User Datagram Protocol (UDP).
- Application Layer
- The Application layer of the Internet Architecture relates to the presentation and application layers of the OSI model. Application programs such as remote terminal protocol (Telnet), File Transfer Protocol (FTP), Simple Mail Transfer Protocol (SMTP) can be invoked to access nodes on the network.
The diagram shows the protocols used in each layer of the Internet Architecture. Each protocol is listed below with it’s respective RFC number.
Internet Protocol (IP) RFC 791
The Internet Protocol (IP) is the basic data transfer method used throughout the Internet. It is responsible for IP Addressing, and performs the routing function, which selects the path to send data to the destination IP Address.
Internet Control Message Protocol (ICMP) RFC 792
The Internet Control Message Protocol (ICMP) is the troubleshooting protocol of TCP/IP. It allows Internet hosts and gateways to report errors through ICMP messages.
Internet Group Management Protocol (IGMP) RFC 1112
The Internet Group Management Protocol (IGMP) is used for multicasting. In multicasting, one source sends a message to a group of subscribers (multicast groups). For multicast delivery to be successful, members must identify themselves and the groups that interest them to local multicast-enables routers.
Address Resolution Protocol (ARP) RFC 826
The Address Resolution Protocol (ARP) is used to translate Internet addresses to physical addresses, such as an ethernet network’s 48-bit physical address.
Reverse Address Resolution Protocol (RARP) RFC 903
The Reverse Address Resolution Protocol (RARP) performs the reverse function of ASRP. It uses a node’s hardware address to request an IP address. RARP is generally used for disl;ess workstations that do not have an IP address.
Transport Control Protocol (TCP) RFC 793
The Transport Control Protocol (TCP) provides session management between the source and the destination systems. TCP is a stateful protocol, i.e. a session must be esatblished prior to transferring data.
User Datagram Protocol (UDP) RFC 768
The User Datagram Protocol (UDP) provides a simple datagram form of communication. One UDP packet is created for each output operation by an application, and a session is not necessary.
Hypertext Transfer Protocol (HTTP) RFC 1945 and 2616
The Hypertext Trnsfer Protocol (HTTP) is used to transport HTML documents (Web Pages) across the Internet. HTTP requires a client program (a browser) on one end and a server on the other, both running TCP/IP.
File Transfer Protocol (FTP) RFC 959
The File Transfer Protocol (FTP) is a system for transferring files between computers on a TCP/IP network. FTP is an efficient and quick way to transfer files as it does not have the overhead of encoding/decoding the data.
Trivial File Transfer Protocol (TFTP) RFC 1350
Trivial File Transfer Protocol (TFTP) is used for initialising diskless systems. It works with the BOOTstrap (BOOTP) protocol.
Introduction to Routing
Routing is a very important function of IP. The device that performs the task is a router, which forwards packets from one physical network to another.
The Internet Layer, or OSI network layer, performs the routing function. A packer, or datagram, contains sufficient information for routing from source host to the destination host. The routing function is not necessarily reliable. The application or transport layer is responsible for reliability, ensuring that the packet reaches the destination.
Routing Process in Internet Architecture
Routing involves two key elements:
- The host must know which router to use for a given destination; the router is determined by the default gateway. The default gateway is the IP address of the router on your local network.
- The router must know where to send the packet; the destination is determined by the router’s routing information table.
Routing Information Tables
A routing information table is a database maintained by the router. The table contains the location of all networks in relation to the router’s location. When a packet arrives at the router, the router examines the packet’s destination network, then checks it’s own routing information table. It determines the next router to send the packet, and forwards the packet to that router, which is considered a hop. Static routers contain routing information tables that must be updated manually, Dynamic routers communicate with other routers to calculate routes.
Routing Information Protocol (RIP)
The Routing Information Protocol (RIP) is commonly implemented on small to medium-sized LANs. RIP maintains only the best route to a destination. Old route information is replaced by new route information, causing network topology changes that are reflected in routing update messages. Routing update messages cause routers to update their tables and propagate the changes. Two versions of RIP are used: RIPv1 (RFC 1058) and RIPv2 (RFC 2453).
Open Shortest Path First (OSPF)
A disadvantage of RIP is that routes are selected on the basis of the closest path (fewest hops) between source and destination systems. No emphasis is placed on factors such as available bandwidth, multiple connection or security.
The Open Shortest Path First (OSPF) routing protocol is an interior gateway routing protocol that overcomes many of RIP’s shortcomings. OSPFv2 (RFC 2328) has recently become and Internet standard protocol.
OSPF contains several practical features:
- Routing Table Updates – Updates take place when necessary, rather than at regular intervals.
- Various types of Service Routing – OSPF makes it possible to install multiple reoutes to a given destination.
- Load Balancing – If multiple routes exist to a destination and all routes cost the same, OSPF distributes traffic evenly over all routes.
- Network areas – OSPF provides the ability to partition a network into areas, allowing growth and organisation.
- Authenticated Exchanges – All exchanges between routers using the OSPF protocol are authenticated.
- Defined Route Support – OSPF allows the definition of host-specific or network-specific routes.
When a packet arrives at a destination host using the IP address, the packet is passed to the transport layer. The transport layer determines which service the packet is using by examining the packet’s destination port number.
TCO and UDP protocol headers contain both the source and destination port numbers. These port numbers are addresses by which processes can be identified. Each port number is a 16-bit integer value that identifies a communication channel to a specific process, For example:
- FTP = port 21
- HTTP = port 80
- DNS = port53
- SMTP = port 25
The standard port assignments are:
|Port Number Range||Description|
|1 to 1023||Well-known (reserved) port numbers|
|1024 to 65535||Registered port numbers|