Cluster Token Registry for Business Continuity

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Business Continuity Clustering (BCC) systems lack managerial functions for individual resources, leading to potential service interruptions when a primary cluster fails, as they rely solely on local decisions based on a 'quorum' concept without ensuring that resources are properly authorized and maintained across geographically dispersed clusters.

Innovation Solution

Implementing a Cluster Token Registry (CTR) process across interconnected nodes, where each node maintains a cluster-wide flag or 'quorum token' that expires if not refreshed periodically, ensuring that services depend on a valid token and preventing unintended resource ownership, using a virtual synchrony protocol for consistent decision-making across the cluster.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If BCC systems rely solely on local decisions based on a 'quorum' concept, then device complexity is reduced, but reliability deteriorates due to potential service interruptions when a primary cluster fails

Engineering Contradiction:
Improvecomplexity of cluster decision-makingVSAvoidservice continuity
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent introduces a Cluster Token Registry (CTR) as an intermediary service that manages cluster-wide tokens and coordinates failover decisions across geographically dispersed clusters. The CTR receives failure notifications, validates prerequisites, and grants takeover tokens to secondary clusters, thereby maintaining reliability while keeping individual node complexity manageable through centralized coordination.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If entire cluster operations are moved to secondary clusters without resource-level authorization, then productivity is maintained during failover, but reliability worsens due to unintended resource ownership

Engineering Contradiction:
Improveservice availability during failoverVSAvoidresource authorization integrity
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent segments cluster resources into individual manageable units, each with its own prerequisites for takeover. Instead of treating the entire cluster as a single entity, the system evaluates and authorizes takeover on a per-resource basis through the CTR, which checks specific prerequisites for each resource before granting tokens. This ensures that only resources meeting their authorization criteria are transferred, preventing unintended resource ownership while maintaining service availability.

Inventive Principle:
Principle #1Segmentation

3Ease of operation

If cluster tokens are granted without periodic verification, then ease of operation is improved, but reliability deteriorates due to token expiration and unauthorized workload assumption

Engineering Contradiction:
Improvesimplicity of token managementVSAvoidauthorized workload assumption
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent implements periodic verification through a heartbeat mechanism where the CTR sends periodic requests to verify token validity. Secondary clusters must respond to these heartbeats to maintain their tokens, and the CTR can revoke tokens if prerequisites are no longer met. This periodic action ensures continuous authorization validation while automating the verification process to maintain ease of operation.

Inventive Principle:
Principle #19Periodic action

Data Source

PatentUS8903917B2System and method for implementing a cluster token registry for business continuity
Publication Date: 2014.12.02 SUSE LLC
  • US8903917B2 patent drawing
  • US8903917B2 patent drawing
  • US8903917B2 patent drawing

AI summary

Method and system for implementing a cluster token registry (“CTR”) process in a cluster comprising a plurality of interconnected nodes each having a CTR associated therewith are described. In one embodiment, the method comprises, responsive to a client signing in at one of the nodes, the client registering a cluster-wide token with the CTR of the node, thereby casting a vote in connection with the token; determining whether prerequisites for granting the token have been met; and, responsive to a determination that the prerequisites have been met, granting the token. The method further comprises periodically determining whether the vote in connection with the token vote has been refreshed by the client and, responsive to a determination that the vote in connection with the token has not been refreshed by the client, de-registering the client.