CTI SOA Clusters for Stateless Event Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing call routing systems require direct TCP/IP connections between agents and peripheral gateway servers, leading to inefficiencies and limitations in scalability and flexibility, as each agent needs a dedicated connection, which is not stateless and does not allow agents to connect to any peripheral gateway server without pre-designation.
Innovation Solution
A Computer Telephony Integration (CTI) Service-Oriented Architecture (SOA) arrangement that groups CTI servers into clusters, enabling stateless and scalable routing of solicited and unsolicited events between agents and peripheral gateway servers, allowing any agent to connect to any peripheral gateway server without a designated TCP/IP connection, using HTTP connections and adapter modules within the clusters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If direct TCP/IP connections are established between each agent and peripheral gateway server, then call routing can be performed, but the system complexity increases and scalability is limited due to the need for dedicated connections
Solution Approach 1:
The patent introduces a message broker as an intermediary component that mediates communication between agents and peripheral gateway servers. Instead of direct peer-to-peer connections, agents publish messages to topics subscribed by the message broker, which then routes messages to appropriate servers. This intermediary architecture eliminates the need for complex direct connection management while maintaining reliable call routing capability.
Solution Approach 2:
The message broker serves multiple functions simultaneously: it acts as a message router, a connection manager, a load balancer, and a message queue. This universal component handles all communication between agents and servers, replacing multiple specialized connection management functions and significantly reducing system complexity.
2Ease of operation
If dedicated TCP/IP connections are assigned to each agent-PG pair, then direct communication is enabled, but the system becomes less flexible and cannot dynamically realign in response to failures
Solution Approach 1:
The message broker implements dynamic routing where subscription relationships can be established and changed at runtime without requiring pre-configured connections. When a peripheral gateway server fails or becomes available, subscribers can dynamically subscribe or unsubscribe from relevant topics, enabling the system to adapt and realign communication paths automatically without manual intervention.
Solution Approach 2:
The communication architecture is segmented into independent publish-subscribe relationships rather than rigid point-to-point connections. This segmentation allows individual subscription relationships to be modified independently, enabling flexible dynamic realignment when system conditions change while maintaining ease of operation through the standardized publish-subscribe interface.
3Productivity
If each agent maintains a direct connection to a designated PG server, then call routing is straightforward, but scalability is limited as the number of connections increases linearly with agents
Solution Approach 1:
The patent merges multiple individual agent- PG connections into a single centralized message broker infrastructure. Instead of having N agents create N separate connections to PG servers (resulting in N connection endpoints), all agents publish to shared topics that the message broker manages, reducing the number of active connection endpoints while maintaining the ability to route calls efficiently to any number of PG servers.
Data Source
AI summary
A system and method for providing a Computer Telephony Integration Service Oriented Architecture is presented. The system and method may include providing one or more computer telephony integration servers in one or more clusters to deliver telephony events between agents and peripheral devices. The events may be solicited or unsolicited. The clusters may be stateless and scalable. The clusters may include dynamic message routing. The systems and methods may implement one or more recovery algorithms if one or more of the telephony integration servers experiences a failure event.