Autonomous System State Tolerance Adjustment for Autonomous Management Systems

Abstract

In general, the techniques of this invention are directed to determining whether a component failure in a distributed computing system is genuine. In particular, embodiments of this invention analyze monitoring data from other application nodes in a distributed computing system to determine whether the component failure is genuine. If the component failure is not genuine, the embodiments may adjust a fault tolerance parameter that caused the component failure to be perceived.

1 . A method comprising:
 receiving monitoring data from a plurality of application nodes interconnected via a communication network;   receiving a failure notification from an autonomic management system when the autonomic management system perceives a failure of a first application node based on a fault tolerance parameter;   executing an autonomous system status manager to analyze the monitoring data from the application nodes to determine whether the perceived failure of the first application node is genuine; and   autonomically adjusting the fault tolerance parameter with the autonomous system status manager when the perceived failure of the first application is determined not to be genuine.   
     
     
         2 . The method of  claim 1 , wherein analyzing the monitoring data comprises autonomously determining whether a degree of computational service utilization exceed a threshold and caused the autonomic management system to incorrectly perceive the failure of the first application node. 
     
     
         3 . The method of  claim 2 ,
 wherein the fault tolerance parameter is a measure of time between attempts to connect to the first node; and   wherein determining whether a service issue caused the autonomic management system to perceive the failure comprises increasing the measure of time between attempts when utilization of the first node is rising.   
     
     
         4 . The method of  claim 1 , wherein adjusting the fault tolerance parameter comprises causing the autonomic management system to ignore monitoring failures during a period of high system service utilization. 
     
     
         5 . The method of  claim 1 , wherein analyzing the monitoring data comprises using historical information and pre-programmed responses to known failure scenarios to analyze the state of the application nodes. 
     
     
         6 . The method of  claim 1 , wherein analyzing the monitoring data comprises using a backward chaining rule set to analyze the monitoring data. 
     
     
         7 . The method of  claim 1 , wherein adjusting the fault tolerance parameter comprises modifying monitoring settings of the first node. 
     
     
         8 . A distributed computing system comprising:
 a plurality of application nodes interconnected via a communications network;   an autonomic management system to provide autonomic control of the application nodes, wherein the autonomic management system monitors the application nodes to perceive a failure of a first node of the application nodes based on a fault tolerance parameter; and   a system status manager to autonomously adjust the fault tolerance parameter based on an analysis of a state of all of the application nodes.   
     
     
         9 . The distributed computing system of  claim 8 , wherein the system status manager adjusts the fault tolerance parameter based when a service issue causes the autonomic management system to detect the failure. 
     
     
         10 . The distributed computing system of  claim 9 ,
 wherein the fault tolerance parameter is a measure of time between attempts to connect to the first node; and   wherein system status manager increases the measure of time between attempts when utilization of the first node is rising.   
     
     
         11 . The distributed computing system of  claim 8 , wherein the system status manager causes the autonomic management system to ignore monitoring failures during a period of high system service utilization. 
     
     
         12 . The distributed computing system of  claim 8 , wherein the system status manager uses historical information and pre-programmed responses to known failure scenarios to analyze the state of all of the application nodes. 
     
     
         13 . The distributed computing system of  claim 8 , wherein the system status manager comprises a monitoring rule engine (MRE) to perform the analysis when autonomic management system detects the failure, wherein the monitoring rule engine comprises a backward chaining rule set. 
     
     
         14 . The distributed computing system of  claim 13 ,
 wherein the MRE generates an action request based on the analysis; and   wherein the system status manager comprises a task manager to institute the action request in the distributed computing system.   
     
     
         15 . The distributed computing system of  claim 14 , wherein the action request includes modification to monitoring settings of the first node. 
     
     
         16 . The distributed computing system of  claim 14 , wherein the action request includes instructions to the autonomic management system to ignore monitoring settings for the first node. 
     
     
         17 . A computer-readable medium comprising instructions, the instruction causing a processor to:
 receive monitoring data from a plurality of application nodes interconnected via a communication network;   receive a failure notification from an autonomic management system when the autonomic management system perceives a failure of a first application node based on a fault tolerance parameter; and   analyze the monitoring data from all of the application nodes to determine whether the perceived failure of the first application node is genuine; and   adjust the fault tolerance parameter when the perceived failure of the first application is not genuine.