Fault-tolerant processing system

• US 5,533,188 A
• Filed: 10/19/1992
• Issued: 07/02/1996
• Est. Priority Date: 10/19/1992
• Status: Expired due to Fees

• Current Assignee: National Aeronautics and Space Administration (Government of the United States of America)
• Original Assignee: United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration
• Inventors: Palumbo, Daniel L.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): HUA, LY
• Application Number:
  US07/963,170
• Time in Patent Office: 1,352 Days
• Field of Search: 371/16.1, 371/68.1-.3, 371/8, 371/11.1-.3, 371/61, 371/1, 371/36, 371/30, 395/575, 395/275, 395/325, 395/182.02, 395/182.08, 370/4, 364/412, 364/419
• US Class Current: 714/4.5
• CPC Class Codes: G06F 11/1679 at clock signal level
• CPC Class Codes: G06F 11/185 and the voting is itself performed redundantly

Abstract

A fault-tolerant, fiber optic interconnect, or backplane, which serves as a via for data transfer between modules. Fault tolerance algorithms are embedded in the backplane by dividing the backplane into a read bus and a write bus and placing a redundancy management unit (RMU) between the read bus and the write bus so that all data transmitted by the write bus is subjected to the fault tolerance algorithms before the data is passed for distribution to the read bus. The RMU provides both backplane control and fault tolerance.

Claims

1. A system for implementing fault-tolerant processing using a fiber optic network, comprising:

a plurality of backplanes including a local backplane and at least one external backplane;

a plurality of processing modules including at least one local processing module and at least one redundant processing module, each said processing module having a bus control, wherein each said local processing module is optically connected to the local backplane and each said redundant processing module is optically connected to one said external backplane, each said external backplane having an optical connection to at least one said redundant processing module, wherein each backplane comprises;

(a) redundancy management means for controlling access to the backplane, for comparing data from each said local processing module with data from each said redundant processing module and for producing corrected data based on the comparison, the redundancy management means of each said local backplane being optically interconnected to each said redundancy management means in each said external backplane;

(b) a local write bus optically connecting each said local processing module to said redundancy management means so that said data from each said local processing module are optically transmittable to said redundancy management means; and

(c) a local read bus optically connecting said redundancy management means to each said local processing module so that said corrected data are optically transmittable to each said local processing module;

wherein each said bus control optically connects said local processing module to the local read bus and the local write bus; and

said redundancy management means includes synchronizing means for synchronizing transmission of said data from each said local processing module through said local write bus and to each said local processing module through said local read bus.

2. A system as recited in claim 1, wherein said redundancy management means further includes:

means for masking failure of each said local processing module and each said redundant processing module.

3. A system as recited in claim 1, wherein said redundancy management means further includes:

means for deskewing said data from each said local processing module and each said redundant processing module.

4. A system as recited in claim 3, wherein the redundancy management means further comprises a clock synchronization unit operatively connected to the deskewing means to compute a suitable clock correction which synchronizes the redundancy management means.

5. A system as recited in claim 1, wherein said synchronizing means includes:

time base generating means for generating a time base, for controlling transmission of said corrected data to each said local processing module through said local read bus based on said time base, and for controlling transmission of said data from each said local processing module to the redundancy management means through said local write bus.

6. A system as recited in claim 4, wherein said synchronizing means further includes:

means for generating a synchronization command and for transmitting said synchronization command to each said local processing module through said local read bus.

7. A system as recited in claim 6, wherein each of the bus controls includes:

means for regenerating said time base based on the timing of said corrected data; and

means for monitoring said local read bus and for setting a first register counter to a global time when said synchronization command is detected.

8. A system as recited in claim 7, wherein each said local processing module further includes:

scheduler means, included in said b,us control and operatively connected to receive said global time from said first register counter, for identifying desired input data within said corrected data based on said global time;

processing means, for processing said desired input data to produce processed data;

storage means, operatively connected to said scheduler means and said processing means for storing desired input data and processed data; and

transmitter means, included in said bus control, for transmitting and time division multiplexing said processed data through said local write bus based on the regenerated time base.

9. A system as recited in claim 8, wherein said redundancy management means includes:

a second register counter setting said global time;

a scheduler processor, operatively connected to receive said global time from said second register counter, having a look-up table of data identification and data local module source versus global time, said scheduler processor generating a data identification and data local module source signal based on said global time; and

a switch, operatively connected to said local write bus and said scheduler processor, selecting said data from each said local processing module based on said data identification and data local module source signal.

10. A system as recited in claim 9, wherein the redundancy management means further comprises:

a distributed system data path controller connected to the scheduler processor and having an input network line for inputting input network data and an output network line for outputting output network data;

means for providing said corrected data to the schedule processor, wherein the scheduler processor selects output network data from the corrected data, wherein said output network data is output to the distributed system data path controller for transmission over the output network line; and

means for providing said input network data from the distributed system data path controller to the switch so that said input network data is compared with input network data from redundant distributed system data path controllers, said redundant distributed system data path controllers being identical to said distributed system data path controller and being connected to redundancy management means in each said external backplane.

11. A system as recited in claim 10 wherein:

at least one of said input network line and said output network line of said distributed system data path controller is connected to a second distributed system data path controller to form a first channel of a distributed system data path, said redundant distributed system data path controllers each having an input network line and an output network line, and each of said redundant distributed system data path controllers being respectively connected to additional distributed system data path controllers to form redundant channels of said distributed system data path.

12. A system as recited in claim 1, wherein each said local processing module includes:

means for multiplexing said data from each said local processing module through said local write bus.

13. A system as recited in claim 1, wherein each backplane further comprises at least two replicated redundancy management means optically connected to the local write bus so that each replicated redundancy management means receives the data from each said local processing module.

14. A system as recited in claim 13, further comprising a plurality of replicated local read buses, wherein each local read bus interconnects a replicated redundancy management means to each said local processing module so that said corrected data from the replicated redundancy management means is transmittable through the replicated local read buses to each said local processing module.

15. A system as recited in claim 14, wherein each said local processing module includes an optical combiner that optically combines said corrected data from said redundancy management means and corrected data from said replicated redundancy management means.