Network Service Level Characterization via Event Log Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The availability of applications in data centers is affected not only by the functioning of the application itself but also by the proper functioning of network devices, and existing solutions do not effectively characterize or improve network service levels, especially in scenarios with redundant devices that may fail together.
Innovation Solution
The method involves characterizing service levels of network devices and applications by processing data center event logs, filtering out redundant events, and analyzing the effectiveness of redundant groups to ensure connectivity, using statistical models to calculate availability and reliability, and identifying potential network changes to meet service level agreements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If redundant network devices are deployed to improve availability, then system reliability improves, but device complexity increases
Solution Approach 1:
The patent segments the network into redundant groups where each group contains multiple network devices that can independently operate. This segmentation allows the system to maintain availability even when individual devices fail, while keeping each individual device's complexity manageable through modular redundancy architecture.
Solution Approach 2:
The patent changes the parameter of device availability from individual component level to group level. By analyzing redundancy group effectiveness as a collective parameter rather than individual device performance, the system can achieve higher overall availability while managing the complexity through statistical modeling and characterization methods.
2Measurement precision
If individual network devices are monitored to assess service levels, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent merges the monitoring of multiple network devices into a unified service level assessment framework. Instead of independently monitoring each device and accumulating complex data, the system combines device-level metrics into redundancy group-level service level characterizations, reducing overall monitoring complexity while maintaining measurement precision.
Solution Approach 2:
The patent introduces service level characterization as an intermediary layer between individual device monitoring and overall system availability assessment. This intermediary framework aggregates and contextualizes device-level data into meaningful service level metrics, simplifying the monitoring architecture while improving measurement accuracy.
3Reliability
If redundancy groups are optimized to improve availability, then system reliability improves, but device complexity increases
Solution Approach 1:
The patent implements feedback mechanisms that continuously monitor and assess redundancy group effectiveness. By analyzing actual service levels and comparing them against target availability metrics, the system can dynamically optimize redundancy configurations without manually increasing complexity, using automated feedback loops to tune redundancy parameters.
Solution Approach 2:
The patent makes redundancy configurations dynamic rather than static. The system can adaptively adjust redundancy group compositions and effectiveness parameters based on real-time service level data, allowing optimization of availability while managing complexity through dynamic reconfiguration rather than fixed complex architectures.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The described implementations relate to processing of electronic data. One implementation is manifest as a system that that can include an event analysis component and one or more processing devices configured to execute the event analysis component. The event analysis component can be configured to obtain events from event logs, the events reflecting failures by one or more network devices in one or more data centers and characterize a service level of an application or a network device based on the events. For example, the event analysis component can be configured to characterize the availability of an application based on one or more network stamps of the application, said network stamps being network devices on which a given application depends to carry application traffic.