Network Analytics Platform for Microservice Performance Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern data centers face challenges in predicting and managing application and network performance due to the complexity introduced by micro-services, virtualization, hybrid clouds, and massively distributed systems, which makes it difficult to assess and improve performance metrics such as latency and availability.
Innovation Solution
An application and network analytics platform that captures telemetry data from servers and network devices, models performance metrics, and predicts how changes affect network performance, enabling real-time insights and automated actions to improve unavailability, load, and latency issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If traditional monitoring methods are used in modern data centers with micro-services and virtualization, then device complexity is reduced, but measurement precision and reliability of performance assessment deteriorate
Solution Approach 1:
The patent segments network traffic into flowlets (sub-flows) based on time gaps and application layer data. This segmentation allows precise tracking of individual application transactions through complex virtualized networks, enabling accurate performance measurement without requiring complex infrastructure changes. Each flowlet is tracked independently to measure application-specific metrics like latency and availability.
Solution Approach 2:
The patent introduces an intermediary analytics platform that sits between network devices and administrators. This platform captures telemetry data from multiple sources (network devices, virtualization layers, applications), correlates flow data with application metadata, and presents unified performance insights. The intermediary handles the complexity of micro-services and virtualization internally while providing simplified precise measurements to users.
2Measurement precision
If comprehensive telemetry data is captured from all network devices and applications, then measurement precision improves, but device complexity and data processing requirements increase
Solution Approach 1:
The patent performs preliminary actions by pre-establishing flow tracking structures and pre-correlating network flow data with application metadata before performance analysis is needed. Application dependency maps and flow-to-application mappings are built in advance, allowing the system to quickly answer performance questions without complex real-time processing. This pre-computation reduces operational complexity while maintaining high measurement precision.
3Measurement precision
If detailed application-level flow tracking is implemented, then measurement precision improves, but loss of time in capturing short-lived micro-service operations increases
Solution Approach 1:
The patent uses periodic sampling at application-layer time gaps (e.g., checking for data between packets in a flow). Instead of continuously monitoring every packet, the system periodically samples flow data at strategic intervals, particularly at application protocol boundaries. This periodic action captures sufficient information about short-lived micro-service operations while minimizing time loss and processing overhead.
Data Source
AI summary
An application and network analytics platform can capture comprehensive telemetry from servers and network devices operating within a network. The platform can discover flows running through the network, applications generating the flows, servers hosting the applications, computing resources provisioned and consumed by the applications, and network topology, among other insights. The platform can generate various models relating one set of application and network performance metrics to another. For example, the platform can model application latency as a function of computing resources provisioned to and/or actually used by the application, its host's total resources, and/or the distance of its host relative to other elements of the network. The platform can change the model by moving, removing, or adding elements to predict how the change affects application and network performance. In some situations, the platform can automatically act on predictions to improve application and network performance.