Method for on-line diagnosis for distributed network systems

• US 5,537,653 A
• Filed: 03/24/1995
• Issued: 07/16/1996
• Est. Priority Date: 12/03/1992
• Status: Expired due to Fees

• Current Assignee: Carnegie Mellon University
• Original Assignee: Carnegie Mellon University
• Inventors: Bianchini, Ronald P. Jr.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Snyder, Glenn
• Application Number:
  US08/409,535
• Time in Patent Office: 480 Days
• Field of Search: 395/575, 395/908, 395/909, 395/916, 395/917, 395/183.01, 395/183.02, 395/183.07, 395/182.02, 395/183.19, 371/15.1, 371/8.2, 371/11.2
• US Class Current: 714/25
• CPC Class Codes: G06F 11/2273 Test methods

Abstract

An on-line adaptive distributed diagnostic method for use in an arbitrary network comprised of N nodes, where N is greater than or equal to 3. The algorithms of the present invention provide on-line diagnosis in the presence of node and link failure. In the network, each node is capable of executing an algorithm used in the diagnostic method to update a packet which communicates between network nodes.

Claims

1. An on-line adaptive distributed diagnostic system for a network of arbitrary topology, comprising:

N nodes where N is greater than or equal to three, each node having a state of either faulty or fault-free;

M communication links where M is greater than or equal to two, each link having a state of either faulty or fault-free and each node being in direct communication with at least one neighboring node through one of the links;

means associated with each node for directly testing a neighboring node, the determination of which neighboring node to test being based on a testing assignment associated with the network;

means associated with each node for indirectly communicating with a non-neighboring node through intervening nodes such that each node can be furnished with diagnostic information about a non-neighboring node in the network;

means associated with each node for coordinating the performance of diagnosis by it with the performance of diagnosis by another node in the network, whereby the need for central synchronization is eliminated;

means for maintaining diagnosis in the absence of a state change by any of the nodes and links; and

means for restoring diagnosis in the presence of a state change by any of the nodes by adaptively replacing the testing assignment with a new independent testing assignment such that every node in the network is tested by a fault-free node.

2. The system of claim 1 wherein each fault-free node is in communication with every other fault-free node.

3. The system of claim 1 wherein the restoring means further restores diagnosis in the presence of a state change by any of the links.

4. The system of claim 1 wherein each node comprises a processor.

5. The system of claim 1 wherein the direct testing means comprises a memory and a computer program stored in the memory.

6. The system of claim 1 wherein the restoring means adaptively changes the testing assignment at each node based on results of tests performed at that node by the direct testing means and based on any diagnostic information furnished to that node from any non-neighboring nodes.

7. The system of claim 1 wherein the indirect communication means comprises message packets.

8. The system of claim 1 wherein the direct testing means periodically tests a neighboring node.

9. The system of claim 1 wherein each node activates the restoring means by propagating a message packet when the node detects a state change of another node that it is directly testing.

10. The system of claim 9 wherein the restoring means is deactivated when a message packet has traversed the network without being updated.