Browse Prior Art Database

Method for detecting conflicting automation policies Disclosure Number: IPCOM000030211D
Original Publication Date: 2004-Aug-02
Included in the Prior Art Database: 2004-Aug-02
Document File: 2 page(s) / 46K

Publishing Venue



This disclosure describes a system that, implemented within an autonomic manager or automation system, detects the deployment of conflicting policy and either notifies of this fact or disables one or more policies based on relative priority. This is achieved without knowledge of the policy content, but based upon action execution patterns that typify conflicting active policies.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Method for detecting conflicting automation policies

An Autonomic Manager provides the means by which policy for resource behaviour is applied without manual intervention. (Note that the scope of this disclosure covers all automation systems even if they are not usually referred to as Autonomic Managers). Autonomic Manager implementations typically provide the hosting environment for programs that apply policy, and a number of services that support the autonomic manager in doing so. The specification of policy, however, is left to the 'user', whether that be a development or services team or really the end user. Conflicting policy specification, as a result of multiple user support or as an unwitting error can negatvely impact the value of an autonomic manager by adding system resource consumption through thrashing (conflicting policies repeatedly overturning each others' actions) and by failing to achieve the goals of either policy.

    This invention reduces the impact of conflicting policy by recognizing patterns that are exhibited by such behaviour and bringing the situation to the attention of the user or shutting off one of the policies based on relative priorities.

    In Autonomic Computing literature, autonomic managers implement the MAPE loop. That is, there is code that implements the following four functions: Monitor (M) - data about a managed resource is collected Analyze (A) - the collected data is analyzed to understand the current situation Plan (P) - given the current situation, what actions, if any, are necessary to correct the situation or to keep things under control Execute (E) - execute the output of the planning exercise.

    If conflicting policies are activated, they will result in competing actions implemented at the execute (E) phase, As the actions are typically names of executables or commands, it is not obvious by looking at the policies themselves that policies may be conflicting, nor will it always be possible for any one programmer to see all of the policies together due to span of control and security definitions.

    This invention identifies conflicting policy at execution time based on recognising patterns of action execution typical of policy conflict. Actions are analyzed, and competing patterns are notified to an administrator or resolved au...