HTTP Connection Redirector for Scalable Server Load Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication systems face challenges in managing large numbers of permanent HTTP connections, particularly in IoT and CPE scenarios, where clients remain connected indefinitely, leading to server resource overload and scalability issues, especially in firewall-protected environments.
Innovation Solution
A system utilizing HTTP/2 and SSE capabilities to maintain permanent HTTP connections, with Connection Redirectors, Connection Handlers, and a Location Registry, allowing servers to proactively send messages and manage millions of clients, while using keepalive mechanisms to monitor connection validity and scalability across geographically distributed servers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If permanent HTTP connections are maintained for millions of clients, then proactive server-to-client communication is enabled, but server resource overload occurs
Solution Approach 1:
The system segments the server cluster into multiple independent HTTP server instances, each managing a subset of permanent connections. This distribution prevents any single server from becoming overloaded while maintaining the ability to proactively communicate with clients. The segmentation is further enhanced by dividing functionality across Connection Handlers and Location Registry components.
Solution Approach 2:
The patent introduces a Location Registry as an intermediary component that maintains the mapping between client identifiers and their assigned HTTP server instances. This mediator enables efficient routing of proactive communication requests without requiring the actual server to store all connection state information, thereby reducing server resource consumption while maintaining adaptability.
2Quantity of substance
If a large cluster of servers is deployed to support millions of permanent connections, then connection capacity increases, but synchronization complexity between instances becomes unmanageable
Solution Approach 1:
The patent extracts the synchronization complexity from the HTTP server instances by moving the connection state management to a separate Location Registry component. This extraction allows multiple server instances to operate independently without requiring complex synchronization mechanisms, as they only need to query the Location Registry for routing information rather than maintain shared state.
Solution Approach 2:
Each HTTP server instance independently queries the Location Registry to determine which connections it should manage, eliminating the need for inter-instance synchronization. The Location Registry itself is self-sufficient in maintaining the connection mapping information, reducing the overall system's synchronization complexity while supporting millions of connections.
3Adaptability or versatility
If permanent connections are maintained for all clients, then server-initiated communication is enabled, but connection management overhead increases significantly
Solution Approach 1:
The system performs preliminary action by establishing permanent HTTP connections and registering them in the Location Registry when clients first connect, rather than creating connections on-demand. This preliminary setup enables efficient server-initiated communication later without the overhead of repeated connection establishment and teardown, reducing long-term management overhead while maintaining adaptability.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A system (100) and method are provided to support a scenario where a cluster of HTTP - Hypertext Transfer Protocol - servers has to accept and maintain permanently open HTTP connections with a large number of client devices (110). The proposed system (100) relies on HTTP/2 and Server-sent Events – SSE - in order to keep connections open and to allow bidirectional message exchange between client (110) and server. It is comprised by following computational entities: at least one connection redirector entity (101), at least two connection handler entities (102), a location registry entity (103) and an execution server entity (104) configured to support an API implementation adapted to provide an interface for external systems (120). The connection between the system (100) and the client device is performed through an internet connection (130).