TCP/IP Networking (I)



TCP/IP Architecture
  • TCP/IP protocol has a four-layer structure linking an application to the physical network.
  • Each layer has its own independent data structures.
  • Conceptually, each layer is speaking directly to its counterpart on the other machine. In this sense, it is ignorant of what goes one after the data is sent.
  • For example, in the Application layer, a NFS Client talks to a NFS Server and knows only the details of the NFS protocol they both use.
  • As data packets are transported from the application to the physical network, each layer adds some control information in the form of a header.
  • Once the packet reaches its destination in the physical network, each layer reads and removes its corresponding header before passing the package up in the stack until it is received by the application.


  • This layer contains all application protocols (often providing user services) that use the Transport layer.
  • Examples of application protocols include FTP, HTTP, DNS, NFS, SMTP, Telnet
  • To send data, the application calls up a Transport layer protocol, such as TCP.
  • Application Layer protocols usually treat transport and lower layer protocols as "black boxes." In this sense, they assume a stable network connection exist across which to communicate.







  • TCP and UDP are the most importan protocols in this layer, delivering data between application and internet layers.
  • TCP provides reliable data delivery service with error detection and error correction. It delivers data received from IP to the correct application (identified by a port number).
  • UPD provides a connectionless delivery service.
  • When called by an application, TCP wraps the data into a TCP packet.
  • A TCP packet (also called TCP segment) contains a TCP header followed by the application data (including header).
  • TCP then hands the packet to IP.
  • TCP keeps track of what data belongs to what process.
  • It is also responsible for ensuring that the packets are delivered with the correct contents and put in the right order before handing them off to the receiving application.

  • The layer above the Network Access layer, and it provides the packet delivery service on which TCP/IP networks are built.
  • It provides a routing mechanism allowing for packets to be transmitted across one or more different networks.
  • The Internet Protocol (IP) runs in this layer and provides a way to transport datagrams across the network.
  • It is a connectionless protocol and does not provide error control, relying on protocols in the other layers to provide error detection and recovery.
  • Source and destination may be in the same or different networks.
  • The IP protocol performs the functions of (a) host addressing and identification, and (b) packet routing (transporting packets from source to destination).
  • After receiving a TCP packet, IP wraps it up and prepends an IP header, creating an IP datagram.
  • Moving the data down the stack, IP hands it off to the hardware driver, that runs in the Network Access Layer.

  • The IP layer has to figure out how to send the packet.
  • Destination on a different physical network ?
    • Then IP needs to find and send it to the appropriate gateway.
  • Destination on the local ethernet network ?
    • IP uses the Address Resolution Protocol (ARP) to determine what Ethernet card's MAC address is associated with the datagram IP address.
  • How does it work?
    • ARP broadcasts an ARP packet across the entire network asking which MAC address belongs to a particular IP address.
    • Although every machines gets this broadcast, only the one out there that matches will respond. This is then stored by the IP layer in its internal ARP table.

You can look at the ARP table at any time by running the command:
jdoe@quark:~$ arp -a
home (194.113.47.147) at 98:0:bd:bd:8c:d2 [ether] on eth0
jdoe@quark:~$ 

  • Protocols in this layer are designed to move packets (IP datagrams) between the internet layer interface of two different hosts on the same physical link.
  • The actual process of moving packets at this level is usually controlled by device drivers of the network cards, which must know the details of the underlying network in order to format the data appropriately.
  • At this level IP addresses are translated to physical addresses used by the network cards (i.e. Media Access Control (MAC) addresses)
  • The network access layer (also called link layer) can be represented by different kinds of physical connections: Ethernet, token-ring, fiber-optics, ISDN, RS-232, etc.

Network Interfaces
  • TCP/IP defines an abstract interface for hardware access.
  • The interface, offering a set of operations that is used to access all types of hardware, hides the implementation details of operations necessary to access each particular equipment. Each vendor is responsible for provinding a driver that translates the commands of the TCP/IP interface to those of the particular piece of hardware.
  • Each networking device has a corresponding interface in the kernel.
  • When configured, each physical device is assigned an interface name.
  • Each interface must also be assigned an IP address. Some interface names include:
    • Ethernet interfaces: eth0, eth1
    • PPP interfaces: ppp0, ppp1
    • FDDI interfaces: fddi0, fddi1
  • A computer having more than one logical or physical network interface is usually called a Multihomed host.

  • An Ethernet network works like a bus system, where a host may send packets (or frames) of up to 1,500 bytes to another host on the same Ethernet.
  • Hosts are identified by a six-byte address hardcoded into the firmware of its Ethernet network interface card (NIC).
  • Ethernet addresses are usually written as a sequence of two-digit hex numbers separated
    by colons, as in aa:bb:cc:dd:ee:ff.

References:
Bautts, Tony, Terry Dawson and Gregor Prudy. 2005. Linux Network Administratos Guide
Hunt, Craig. 2002. TCP/IP Network Administration

No comments:

Post a Comment