Per-Process Server Redundancy via Active-Standby Notification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing server redundancy systems, such as those using VRRP, require both active and standby servers to be on the same network segment, limiting process-level redundancy and failing to provide real-time status recognition between processes, leading to inefficient failure handling and reduced system robustness.
Innovation Solution
Implementing a system where multiple servers execute processes in parallel, with one server acting as an active process and another as a standby, allowing for periodic operation notifications and mode switching on a per-process basis, and utilizing separate communication networks with VPN connections for redundancy and fault tolerance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If VRRP protocol is used for server redundancy, then active-standby server configuration is achieved, but all processes must be switched together causing inefficient failure handling
Solution Approach 1:
The patent segments the server redundancy from process-level redundancy. Instead of treating the entire server as a single redundancy unit, it enables individual processes to have their own active-standby configuration. This allows selective failure handling where only failed processes trigger switching, while other processes continue operating normally.
Solution Approach 2:
The patent introduces dynamic process status recognition where processes can detect whether their counterpart is operational through inter-process communication or management server queries. This dynamic awareness enables processes to make informed decisions about switching based on real-time status information rather than static server-level redundancy.
2Loss of information
If active server transmits advertisements to standby server, then active server status is notified, but standby server cannot recognize active server state
Solution Approach 1:
The patent implements feedback mechanisms where the active process sends operation notifications to the standby process, and the standby process can query the management server to confirm the active process's status. This bidirectional information flow ensures both servers maintain accurate awareness of each other's operational state.
Solution Approach 2:
The management server acts as an intermediary that maintains status information about all processes and provides this information to both active and standby processes. This centralizes status management and ensures accurate bidirectional communication without requiring direct complex interaction between processes.
3Device complexity
If servers are installed on same network segment, then VRRP communication is simplified, but network failure tolerance is reduced
Solution Approach 1:
The patent allows servers to be deployed on separate network segments or different physical networks while maintaining redundancy functionality. The active and standby servers can communicate through VPNs or other virtual connections, separating the redundancy mechanism from the physical network topology.
Solution Approach 2:
The management server serves as an intermediary that enables status communication between servers on different network segments. It queries process status from the active server and notifies the standby server, allowing distributed servers to maintain coordination without direct peer-to-peer communication.
Data Source
AI summary
A server system including a plurality of servers achieves redundancy on a per process basis. A first server and a second server execute a plurality of processes in parallel. Processes (B1-1, B2-1, and Br-1) in the first server are active processes, and processes (B1-2, B2-2, and Br-2) in the second server are standby processes. The active process periodically transmits an operation notification notifying normal operation, to the standby process. When no longer receiving an operation notification from the active process, the standby process switches an operation mode of the standby process from a standby mode to an active mode and starts provision of service of the process. The process switching is performed on a per process basis instead of on a per server basis.