Chapter 1 - Introduction and Description
11
The session layer corrdinates the exhange of information between systems by using conversational
techniques, or dialogues. Dialogues are not always required, but some applications may require a way of
knowing where to restart the transmission of data if a connection is temporarily lost, or may require a periodic
dialog to indicate the end of one data set and the start of a new one.
Presentation layer protocols are part of the operating system and the application the user runs on a
workstation. Information is formatted for display or printing in this layer. Codes within the data, such as tabs
or special graphics sequences, are interpreted. Data encryption and translation of other character sets are
also handled in this layer.
Applications access the underlying network services using defined procedures in this layer. The application
layer is used to define a range of applications that handle file transfers, terminal sessions, and message
exchanges. Data starts at the application and presentation layers, where a user works with a network
application, such as an electronic mail program. Requests for services are passed through the presentation
layer to the session layer which begins the process of packetizing the information. A connection-oriented
communication session may be opended between the two systems to provide reliable transmissions. Once
the session is established, protocol layers begin exchanging information as appropriate.
1.3.1 Packet
A packet is a package of data that is exchanged between devices over a data communications link. Data
exchanged between devices may include messages and commands, control codes for managing the session,
and data. Information is placed in packets by various communications subsystems of the transmitting system,
then framed into a serial bit-stream and sent across the communications link. One of the main reasons for
packetizing and framing information is that errors on the communication link only affect a small, discernible
part of the transmission, which is easily retransmitted.
The process starts at the application layer. An application in one computer has some information to send to
another computer. The information is sent down through the protocol stack, across the wire, and up through
the protocol stack of the other computer. This information takes the fom of a protocol data unit (PDU). As the
PDU moves through the layers of the sending computer, each layer attaches specific information to the PDU
that is relevant to the protocols in that layer. This attached information is destined for the peer layer in the
receiving computer. For example, the transport layer in the sending computer adds a sequence number to the
PDU. This sequence number is read by the transport layer in the receiving computer to resequence the
packets.
The communication protocol defines the packet structure and the networking system in use and defines the
frame structure for bit-stream transmission. Each protocol layer attaches information destined for its peer
layer in the other computer. When the PDU reaches the physical layer, it is transmitted as a stream of bits.
On copper wire, the bit-streams take the form of voltage level changes that represent binary ones and zeros.
The data-link layer places the PDU in one or more frames and uses a media access method (such as CSMA/
CD or token) to gain access to the transmission media. The PDU might be subdivided into a number of
separately addressed frames. The frame format is defined by the network in use. For example, Ehternet
defines frames that can hold up to 1,500 bytes of packet data.
In general, a packet is a collection of information that contains data (payload) and headers (pilot). Headers
include the source and destination address as well as control information to handle errors and keep packets
flowing properly. Each packet is a separate block of information that can have a different destination address,
and in some cases, different sizes. A typical packet holds 512 bytes of information, so it takes many packets
to transfer a large file over a network.
As packets traverse a network, the addressing information contained in them is used by bridges and routers to
direct packets to their destination, or keep them off of networks where they don't belong. Routers direct
packets along a specific path that has been predetermined as the best route to the packets' destination.
Routers use algorithms to determine efficient paths in conjunction with other routers on the network.
Administrators can also manually configure routers, based on the cost or speed of routes. Information can be
transmitted to another system using either connection-oriented or connectionless methods. A connectionless
session is more efficient for short, bursty transmissions since a session setup is not required, but the end
system may need to perform more work to resequence packets that arrive out of order and check for lost
packets.