Specifications

Ethernet Features
2-6 Introduction
Ring Topology
A ring topology is a point-to-point topology in which the network devices
are connected, device to device, in an unbroken circle. Each signal to be
transmitted on the network must be processed by each station on the ring
before it is passed (or repeated) to the next station.
General Characteristics
Ring topologies commonly use an access method that is called token
passing. No station may talk unless that station has a free token, or
specialized signal code designated to determine which station on the
network is allowed to transmit. The token is passed from station to station
on the network along with the data being transmitted until it is released
by the receiving station.
Ring topologies can be complex in nature. They are easy to expand but
may involve calculations of cable lengths to keep the network within
specification. Most modern ring topologies resemble a physical star but
careful examination will reveal a logical ring cabled in a star
configuration. The use of networking hardware such as modular hubs
takes care of maintaining the logical ring in the wiring closet.
Vulnerability
Adding or removing network stations is simple and can, in most cases, be
done while the network is in operation. High level software takes care of
the recognition of problem nodes and also, in most cases, will remove the
problem nodes from the network and automatically reconfigure the ring.
Star Topology
A star topology is a point-to-point network in which the network devices
are connected through a central concentrator or controller. Two types of
access methods are employed: polling and contention.
Polling Star Topology
On a polling network, devices cannot talk or send messages unless they
are given permission (or polled) by a central computer or controller. A
device must wait to transmit until the controller asks for the information.
Performance of a polled network is dependent on the performance of the
controller and the number of devices attached to the controller.
Failure of the controller in a star topology network will bring the network
down. Failure of an individual node typically will not affect the remainder
of the network.