Browse Prior Art Database

Use of Centrality Metrics to Determine Recovery Order Through the Evaulation of Dependency Data

IP.com Disclosure Number: IPCOM000236813D
Publication Date: 2014-May-16
Document File: 1 page(s) / 21K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for establishing the order in which datacenter components are chosen for recovery following a disaster. The method treats the entire set of dependency relationships as a directed graph and uses centrality algorithms to return centrality scores for each component; components with the highest scores are given priority for recovery.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 01 of 1

Use of Centrality Metrics to Determine Recovery Order Through the Evaulation of Dependency Data

Determining the order of component recovery in a disaster recovery (DR) situation is a critical precursor to executing a successful recovery, but is a problem without a determinate solution. Various techniques, from subject-matter expert (SME) interviews to application dependency mapping software, can identify the datacenter components upon which an organization depends; therefore, those components must be recovered before others are. Unfortunately, examining the full mesh of dependency relationships for an environment of only moderate complexity reveals a number of "circular" or "cross-linked" situations in which the dependencies do not imply a clear recovery order. As an example (grossly simplified), if server Z must precede server Y, and server Y must precede X, then it is clear that Z should be recovered first. However, if A should precede B, and B should precede C, but server C in turn should precede A, then the best recovery order is not clear.

The novel solution is a recovery method that treats the entire set of dependency (and therefore recovery precedence) relationships as a directed graph in which datacenter components are treated as nodes and dependency relationships are treated as vertices. Calculate a directed eigenvector centrality metric for each node in the graph. An order of recovery that places the components with the highest centrality scores earlie...