Multiplex communications – Fault recovery – Bypass an inoperative channel
Reexamination Certificate
1998-03-16
2001-07-03
Olms, Douglas (Department: 2661)
Multiplex communications
Fault recovery
Bypass an inoperative channel
C370S216000
Reexamination Certificate
active
06256293
ABSTRACT:
TECHNICAL FIELD
The present invention relates generally to point-to-point transmission links employing multiple channels. More specifically, it relates to providing fault tolerance on optical wavelength-division-multiplexed multichannel links.
DESCRIPTION OF THE PRIOR ART
A multichannel link comprises a number of channels, say N, between two sites. These channels may be transmitted separately (for example over parallel wires or fiber cables) or multiplexed on to one or a small number of wires or fibers using time or wavelength division multiplexing. Usually these links are realized in the form of line cards, one for each channel. A line card can be any device that provides an interface between the I/O for the channel and the transmission medium. Line cards may fail. In many applications it may be desirable to reroute inputs/outputs (I/Os) of failed line cards on to other line cards (or channels).
One example is the IBM MuxMaster system [1]. This system multiplexes up to ten full-duplex channels on to a single transmission link. Typically, fault management is done by having a fully redundant configuration [2]. For each operating link with 10 channels, an additional link with 10 channels is used as a backup. Switching is provided outside the system so that if a card (channel) on the operating link fails, the I/O for that card is switched to the card on the backup link. However this is an expensive solution since it doubles the hardware needed. In cases where the failure of cards is rare and it is unlikely that more than a few cards fail at any given time, it is desirable to minimize the number of additional spare cards.
U.S. Pat. No. 5,313,456 “Data Link Protecting System”
This patent considers a transmission system comprising of two ends A and B with N transmission lines between them. Each transmission line can carry data signals and an additional low-speed control signal. At any given time there is one control signal that is carried on one of the N transmission lines. At each end A and B this control signal is connected via a 1×N switch to any of the N transmission lines. If the current transmission line fails, the patent teaches how to then switch the control signal onto another unused transmission line using the 1×N switches. Thus this patent provides for rerouting 1 input signal to a spare transmission line in the event of a failure using two 1×N switches. This patent requires complex and expensive 1×N switching components when N is large (e.g., N=100).
U.S. Pat. No. 5,434,691 “Communications System Having Optical Transmission Line . . . ”
This patent provides for rerouting inputs on a failed transmission channel to a spare channel, like the present invention but uses more inefficient methods. The first method taught assumes a 1×N switch on the spare line card which is connected to each of N other line cards. If one of the latter line cards fails the input of that card is then routed on to the spare transmission channel. This patent requires complex and expensive 1×N switching components when N is large (e.g., N=100). The second and third methods taught in this patent connect the link cards in a line or ring network, wherein card i is connected to i−1 and i+1, as in our invention. If a card fails, its input is rerouted through the network to the spare card without affecting the other channels. In the worst case this can take up to N additional “hops”, since the input may have to be routed through N−1 other cards before it reaches the spare card. Such a large number of hops can adversely impact signal quality.
U.S. Pat. No. 4,365,247 “End Station Digital Transmission Link Relieving System”
This patent essentially teaches a specific implementation of the first method of U.S. Pat. No. 5,434,691 wherein a 1×N switch is used on the spare line card to connect it directly to the N other link cards. Again, this patent requires complex and expensive 1×N switching components when N is large (e.g., N=100).
U.S. Pat. No. 5,113,391 “Intelligent Channel Unit”
This patent teaches cascading “intelligent channel units” (ICUs) so that ICU i is connected to ICU i−1 and ICU i+1. If ICU i fails, the patent teaches a method to bypass ICU i so that ICU i−1 can be directly connected to ICU i+1, while ICU i is repaired. Here, the I/O of ICU (line card) i is lost. This patent does not teach how to connect link cards so that if two cards i and j fail simultaneously, they can both be routed to different designated spare cards.
U.S. Pat. No. 5,214,692 “Bypass for Telephone Switching System”
This patent considers a system wherein N digital lines arrive and pass through a “main distribution frame” (MDF) to a “automatic call distributor” (ACD) which contains a N×N switch that connects each digital line to a digital telephone at the output of the switch. In the event of the switch in the ACD failing, the patent teaches two techniques (a) to have each digital telephone connected directly to a digital line (no switching capability provided) bypassing the failed switch, or (b) the digital telephones are connected to another switch in the MDF which is used to connect them to the digital lines if the switch in the ACD fails. This patent does not thus teach the idea of having line cards (or other entities) connected together so that if one or two of these entities fails their inputs can be routed to designated spare cards.
U.S. Pat. No. 5,331,631 “N+K Sparing in a telecommunications Switching Environment”
This patent teaches how to provide fault-tolerance within a multistage packet-switched interconnection network. It does not teach how to reroute inputs on failed line cards to other spare cards using limited switching and small number of additional hops.
SUMMARY OF THE INVENTION
The present invention describes methods for rerouting channels on failed line cards to spare line cards. It uses very limited switching in combination with carefully designed interconnection patterns to achieve rerouting of multiple failed channels to spare line cards by going through a minimal number of additional switching components. The invention provides some switching in the line cards to allow rerouting of the I/O channels in the event of line cards failing. While this invention is applicable to any multichannel system, it has particular advantages when using wavelength division multiplexing, where the number of channels per fiber is not very large and the probability for a line card to fail is relatively high.
Two types of rerouting are considered: (1) where we are allowed to reroute all I/Os in the event of a line card failure (full rerouting), and (2) where we are allowed to reroute only the I/O corresponding to the failed lie card (limited rerouting).
Two different types of switching arrangements are considered. In the first arrangement, referred to as limited switching, each line card is provided with a switch that interconnects the I/O either to the link port (for transmission on the link) or to one of k other local ports. It is also possible to interconnect a local port with the link port. Each port can be interconnected with only one other port at any given time. Local ports of different line cards may be “hardwired” together. The term “wire” shall refer to any physical connection such as a copper wire of optical fiber between two elements being “hardwired” to each other. This allows an I/O of a line card to be switched to the I/Os and link ports of other line cards.
With the first arrangement one line card is designated as a spare card. If another line card fails then the invention proposes a method to reroute all the I/Os on to other channels (full rerouting). Each I/O is routed through only one additional line card. The number of additional cards (or “hops”) a signal is routed through is important since the degradation in signal quality is approximately proportional to this number. Thus this rerouting method minimizes the degradation in signal quality. A similar method is proposed
Gerstel Ornan Alexander
Hall William Eric
Ramaswami Rajiv
Sasaki Galen Hajime
McAndrews Held & Malloy Ltd.
Olms Douglas
Pizarro Ricardo M.
Tellabs Operations Inc.
LandOfFree
Fault management in a multichannel transmission system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Fault management in a multichannel transmission system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Fault management in a multichannel transmission system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2469185