=[ x25-box.com ]=

the network is to receive the packet, look at the destination address,
and from its knowledge of the network send the packet to the next
appropriate component. See Fig. 1.4.

Fig. 1.4 A packet switching network

The packet has destination address C. Component A therefore must
route the packet to component B. Component B has a choice; either of
the onward routes is acceptable. The upper route looks preferable because
there are fewer components; therefore the packet will suffer less delay.
However, if the upper route is already being heavily used, or if it has a
noisy, error-prone component, then the lower route is preferable.
Whichever is chosen and for whatever reason the packet will arrive at
component C as requested. When it does arrive, the receiver can extract
the data and can determine the sender.

It is important to note that component C is a part of the network. It
may not be a user or a service. For example if the packet is being routed
to a screen, then only the data should appear, not the addresses. The
network must deliver information appropriate to the end point, which is
why both the service and the user of it must be considered users of the
network.

When a user obtains access via the network to an accounting package
in a computer, then neither the user nor the accounting package is
concerned with addresses or the routing of packets. All of these are
network functions, and both the user and the computer running the
package are users of the network.

1.4 The layered network model

The above explanation of packet switching is simplistic because it does
not address important issues such as the following:

• What is the format of the packet? How long are the fields in the
packet and how are they encoded—Binary, ASCII, EBCDIC, etc?

НАЗАД