Browse Prior Art Database

Fault Recovery in Multichannel Optical Links

IP.com Disclosure Number: IPCOM000111510D
Original Publication Date: 1994-Feb-01
Included in the Prior Art Database: 2005-Mar-26
Document File: 2 page(s) / 70K

Publishing Venue

IBM

Related People

Italiano, GF: AUTHOR [+2]

Abstract

Future high-speed optical links will use wavelength-division-multiplexing, where multiple high-speed data streams (typically up to 1 Gb/s each) will be sent over the link using different wavelengths of light. It is important to recover efficiently from faults where either the link or one of the nodes at the end of the link fails.

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

Fault Recovery in Multichannel Optical Links

      Future high-speed optical links will use
wavelength-division-multiplexing, where multiple high-speed data
streams (typically up to 1 Gb/s each) will be sent over the link
using different wavelengths of light.  It is important to recover
efficiently from faults where either the link or one of the nodes at
the end of the link fails.

      In the invention, it is assumed in the event of link failure
that there is another path using which the two nodes may communicate.
It is further assumed that either node can be put into loop-back
mode, a mode wherein all the signals on the link from the other node
are reflected back to that node on the link.

      Periodically, the nodes A and B exchange test messages to
determine whether the link is up or down, as in other links.  If no
response is received from node B, then node A knows that either the
link is down or that node B is down.  If node B is up (i.e., the link
is down), then it goes into loop-back mode, where all the energy from
node A is reflected back to that node.  Node A continues to test the
link periodically by sending a test message on one of the channels
and checking if it is reflected back.  If it is reflected back, then
the link is considered to have come up.  Node A then informs node B
using the separate communication path (assumed to be available) that
the link is up again.  Node B then comes out of loop-back mode and
resumes normal operation.

      If node B is down, then when it comes back up, it goes into
loop-back mode.  Node A continues to test the link periodically by
sending a test message on one of the channels and checking if it is
reflected back.  If it is reflected back, then the link is considered
to have come up.  Node A then informs node B using the separate
communication path (assumed to be available) that the link is up
again.  Node B then comes out of loop-back mode and resumes normal
op...