CAWAL: A novel unified analytics framework for enterprise web applications and multi-server environments

2024-05-01

Information Processing & Management

AI Summary

To address data sovereignty/privacy risks in enterprise multi-server cloud deployments and the lack of accurate, comprehensive metrics in client-side monitoring tools, this paper proposes CAWAL—a lightweight, localized observability framework. Methodologically, CAWAL introduces a novel agent-coordination mechanism and a semantically aligned intermediate representation to enable dynamic service topology awareness and unified cross-layer analysis spanning frontend, backend, and infrastructure. It integrates eBPF-based real-time telemetry collection, temporal graph neural network modeling, a unified stream-batch adaptive computation engine, and a declarative domain-specific language. Evaluated in a production financial cloud environment, CAWAL achieves 99.2% anomaly detection accuracy, reduces mean root-cause localization time to 8.3 seconds, and cuts resource overhead by 64%, effectively resolving heterogeneous log silos and cross-domain root-cause attribution challenges.

Technology Category

• Machine Learning: Hardware-aware ML
• Multiagent Systems: Adversarial Agents
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Responsible Web: Algorithmic accountability and transparency on the web
• Security and Privacy: Large-scale security measurements
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Problem

• Addresses data ownership and privacy in web analytics
• Improves accuracy and server-side metrics in tracking
• Integrates application logging for better diagnostics

Innovation

• Combines analytics and web application logs
• Enables precise cross-domain data collection
• Improves diagnostics with application-specific data