Distributed Server Failure Prevention via Shared Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed systems, especially those using cloud computing, failures such as deadlocks, excessive memory consumption, and excessive CPU usage can occur due to defects in application programs, leading to system-wide failures and increased operational costs as the number of servers increases, with existing technologies failing to prevent these failures without imposing a significant load on system operators.
Innovation Solution
A distributed system architecture where first and second servers, capable of executing the same application, generate and share failure information to perform failure prevention processing, allowing the second server to prevent failures based on identified causes, and a distributed management server transmits failure information to enable proactive failure prevention across the system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the number of servers is increased to distribute load, then system capacity and availability are improved, but the likelihood of failures occurring and the operational cost increase
Solution Approach 1:
The patent implements preliminary action by having servers proactively share failure information and perform failure prevention processing before failures occur. When a failure is detected in one server, the failure information is immediately shared with other servers, which then take preventive measures to avoid similar failures, thus resolving the contradiction by maintaining reliability while allowing system expansion.
Solution Approach 2:
The patent employs feedback mechanisms where failure information from one server is fed back to other servers in the distributed system. This feedback loop enables continuous improvement of system reliability as each server learns from failures experienced by others, allowing the system to scale without proportionally increasing failure likelihood.
2Reliability
If application program is modified to fix defects, then failure prevention is improved, but system downtime and operational disruption increase
Solution Approach 1:
The patent implements self-service by enabling servers to automatically perform failure prevention processing based on shared failure information, without requiring manual application program modifications. The system autonomously adjusts its operation to prevent failures, thereby maintaining reliability while avoiding the downtime associated with manual updates.
Solution Approach 2:
The system performs preliminary failure prevention actions by analyzing shared failure information and proactively adjusting operations before failures occur. This allows the system to prevent failures through operational adjustments rather than requiring application program modifications, thus avoiding system downtime.
3Reliability
If temporary measures such as periodical restart are taken, then failure recurrence is prevented, but operational cost and system stability worsen
Solution Approach 1:
The patent uses feedback mechanisms where failure information is continuously shared across the distributed system, enabling servers to learn from past failures and implement permanent preventive measures. This eliminates the need for recurring temporary measures like periodical restarts, thereby reducing operational costs while maintaining failure recurrence prevention.
Solution Approach 2:
The system performs preliminary analysis of failure information and implements preventive measures before failures recur. By proactively addressing root causes through shared learning, the system prevents failure recurrence without requiring repeated temporary measures, thus reducing operational costs associated with frequent restarts and maintenance.
Data Source
AI summary
A distributed system according to an exemplary embodiment includes first and second servers capable of executing the same application, wherein when a failure occurs in the application in the first server, the first server generates failure information identifying a cause of the failure in the application, and the second server performs failure prevention processing which is determined based on the failure information and intended to prevent a failure in the application.