Browse Prior Art Database

Performance Enhancement Via Multiple Logical Subnetworks in An Optical Passive Star Network

IP.com Disclosure Number: IPCOM000101380D
Original Publication Date: 1990-Aug-01
Included in the Prior Art Database: 2005-Mar-16
Document File: 3 page(s) / 106K

Publishing Venue

IBM

Related People

Cheng, T: AUTHOR [+3]

Abstract

Disclosed is a method to enhance the performance of an optical passive-star computer network where a large number of stations are interconnected through a single optical star coupler. The method divides the network into multiple logical subnetworks by the use of multiple optical wavelengths.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Performance Enhancement Via Multiple Logical Subnetworks in An Optical Passive Star Network

       Disclosed is a method to enhance the performance of an
optical passive-star computer network where a large number of
stations are interconnected through a single optical star coupler.
The method divides the network into multiple logical subnetworks by
the use of multiple optical wavelengths.

      Optical passive-star networks are a reliable and inexpensive
means for implementation of multiprocessor networks, I/O buses, and
workstation clusters.  Although the optical medium used in an optical
passive-star network allows bandwidths of several gigabits/second,
there are many practical limitations to exploiting this bandwidth.
One of them is the overhead introduced by the medium-access protocol.
Two of the most widely-used protocols for control of access to the
medium are CSMA/CD (Carrier-Sense Multiple-Access with
Collision-Detection) and the token bus. Both these protocols limit
the utilization of the medium when the number of stations is large.
In the case of CSMA/CD, frequent collisions between stations degrade
the performance.  In the case of token-bus, the performance is
limited by the token-rotation time, which increases with the number
of stations.  The disclosed method is meant to alleviate this
medium-access overhead in networks with a large number of stations.

      The method creates multiple logical subnetworks in the same
physical network by the use of wavelength-division multiplexing.
Each subnetwork is operated with its own medium-access protocol.  The
implementation described below uses two wavelengths.  It can easily
be generalized to more than two wavelengths.

      The implementation for two wavelengths is illustrated in Figure
1.  The stations are divided into two groups such that the
intra-group access-probabilities of the stations are much higher than
the inter- group access-probabilities. Each group is then assigned to
a distinct logical subnetwork within the same physical subnetwork.
The two subnetworks are created by the use of two wavelengths W1 and
W2. In the figure, each station is equipped with two
transmitter-receiver pairs, one for each wavelength.  This allows the
subnetworks to be configured in a fle...