Stateful Cluster Nodes for Unified Communication Scalability

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional unified communication systems face challenges with scalability, redundancy, and complexity, leading to high costs and limitations in user mobility, due to their stateless configurations and hardware-intensive requirements, which are undesirable in modern IT environments.

Innovation Solution

A stateful cluster node system is introduced, where nodes are clustered and communicate with a coordinator and an external message bus, maintaining session information and allowing for efficient routing of events based on customer association, reducing network traffic and processing needs, and enabling seamless failover without disrupting service.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If traditional stateless micro-services architecture is used, then system scalability is improved, but session information management becomes complex and performance drops

Engineering Contradiction:
Improvesystem scalabilityVSAvoidperformance
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The system is divided into multiple independent cluster nodes, each capable of maintaining session information locally. This segmentation allows the system to scale horizontally by adding more nodes while each node independently manages its own session data, avoiding the performance degradation associated with centralized session management in traditional stateless architectures.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Session information is copied and maintained locally at each cluster node rather than being constantly retrieved from a centralized source. This copying mechanism enables nodes to process requests independently using their local session data, significantly improving performance while maintaining scalability through the distributed nature of the copied session information.

Inventive Principle:
Principle #26Copying

2Reliability

If discrete hardware nodes are used in legacy systems, then system redundancy is improved, but system complexity and hardware requirements increase

Engineering Contradiction:
Improvesystem redundancyVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

Multiple discrete hardware nodes are merged into a unified cluster node system where nodes communicate through standardized interfaces and protocols. This merging reduces overall system complexity by providing a consistent architecture for redundancy while maintaining the reliability benefits of having multiple physical nodes through the cluster coordination mechanism.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

Each cluster node is designed with universal functionality to perform multiple roles including session management, event processing, and coordination communication. This multi-functionality reduces the need for specialized hardware components, thereby decreasing device complexity while maintaining system redundancy through the versatile nature of each node.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Ease of manufacture

If stateless micro-services architecture is implemented, then system maintenance cost is reduced, but code rewrite requirements increase

Engineering Contradiction:
Improvesystem maintenance costVSAvoidcode rewrite requirements
Core Design Contradiction:
Ease of manufactureVSDevice complexity

Solution Approach 1:

Session information is maintained proactively at each cluster node before it is needed for processing. This preliminary action of pre-loading and caching session data eliminates the need for complex state management logic in the code, reducing maintenance requirements while avoiding the need for extensive code rewrites by working with an enhanced rather than reduced state model.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3457668B1Clustering in unified communication and collaboration services
Publication Date: 2022.08.10 MITEL NETWORKS INC
  • EP3457668B1 patent drawingFigure 1
  • EP3457668B1 patent drawingFigure 2~3
  • EP3457668B1 patent drawingFigure 4

AI summary

A method for telecommunications is disclosed herein including a first node receiving a plurality of events. A first state machine in the first node generates a current state of the first state machine by processing the plurality of events. The first node publishes the plurality of events to a persistent event log. A second node subscribes to the persistent event log. A second state machine in the second node generates the current state of the first state machine on the second state machine by processing the plurality of events from the persistent event log.