Secure MNO API Gateway with GLM-4 & Kubeflow Threat Ops
Given the challenges MNOs face with Open Gateway/CAMARA initiatives, this challenge tasks developers with building a secure and scalable API Gateway. This gateway will simulate exposing core MNO capabilities (e.g., location services, quality-of-service on-demand) through well-defined APIs. A critical component of this challenge is the integration of an AI-powered threat detection system. This system will analyze API traffic in real-time to identify anomalies such as API abuse, Distributed Denial of Service (DDoS) attacks, or unauthorized access patterns. The threat detection model will be managed and deployed using Kubeflow, leveraging GLM-4 via OpenRouter for advanced contextual analysis, incident reporting, and intelligent mitigation suggestions. The goal is to demonstrate a robust, secure, and AI-augmented network API infrastructure.
What you are building
The core problem, expected build, and operating context for this challenge.
Design and implement a secure API gateway for simulated MNO services, integrating an AI model for real-time threat detection and contextual analysis using GLM-4 via OpenRouter, with MLOps managed by Kubeflow.
How work is evaluated
The challenge will be evaluated by simulating various API traffic patterns, including legitimate usage, stress tests, and malicious attacks, against the participant's deployed API Gateway and threat detection system. The effectiveness of the gateway's security mechanisms and the accuracy and responsiveness of the AI threat detection will be assessed.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Master the design principles of API Gateways (e.g., request routing, throttling, authentication, caching) using a framework like Kong Gateway, Apigee, or by building a custom gateway with FastAPI/Spring Cloud Gateway.
Implement robust security measures for API endpoints, including OAuth2/OpenID Connect for authentication and fine-grained authorization (e.g., JWT validation, role-based access control).
Develop a simulated MNO backend service in Python (e.g., using Flask/FastAPI) that provides basic telecom capabilities (e.g., simulated location data, bandwidth on demand) to be exposed via the gateway.
Design and implement a data ingestion pipeline to collect API gateway logs (request headers, body, IP, timestamps, user-agent) and traffic metrics, pushing them to a message queue like Kafka or directly to a logging service.
Build a machine learning model using TensorFlow/PyTorch or scikit-learn (e.g., time-series anomaly detection, classification for attack types) to analyze the streaming API traffic and detect suspicious patterns indicative of DDoS, API abuse, or unauthorized access attempts.
Orchestrate the entire ML pipeline, including data preprocessing, model training, validation, and deployment, using Kubeflow Pipelines and Kubeflow Serving, ensuring continuous integration and delivery of the threat detection model.
Integrate the GLM-4 API (via OpenRouter) into the threat detection system to provide contextual analysis of detected anomalies, generating detailed incident reports, potential root causes, and recommended mitigation strategies based on API logs and security best practices.
Implement a real-time alerting mechanism (e.g., email, Slack notification) that triggers when an anomaly is detected, enriching the alert with the GLM-4 generated insights.
Reference links and supporting material
Sample data for 3 tasks
Challenge setup
The challenge will be evaluated by simulating various API traffic patterns, including legitimate usage, stress tests, and malicious attacks, against the participant's deployed API Gateway and threat detection system....
Challenge input
{ "gateway_config_url": "https://github.com/user/repo/gateway_config.yaml" }
Agent execution
No agent tool is listed for this challenge.
Evaluated output
URL of the deployed gateway.
- API Gateway and MNO services deploy successfully and are accessible.
- Gateway correctly enforces authentication/authorization for protected endpoints.
- Threat detection system reports no threats during periods of normal, legitimate traffic.
- Threat Detection Accuracy target: 0.95
- Docker execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- API Gateway and MNO services deploy successfully and are accessible.
- Gateway correctly enforces authentication/authorization for protected endpoints.
- Threat detection system reports no threats during periods of normal, legitimate traffic.
- System detects simulated DDoS attacks.
- System detects attempts at unauthorized API access.
- Kubeflow pipeline for model updates/deployment is functional.
Metrics
- Threat_Detection_Accuracy · Target 0.95
- Analysis_Relevance · Target 0.85
- Gateway_Throughput_Under_Load · Target 500
Workflow tasks
- Task 1
Input: { "gateway_config_url": "https://github.com/user/repo/gateway_config.yaml" }
Output: URL of the deployed gateway.
Execution harness: Docker
- Task 2
Input: { "gateway_url": "http://localhost:8000", "traffic_profile": "ddos_simulation" }
Output: Stream of API logs to a specified Kafka topic or log file.
Execution harness: Python
- Task 3
Input: { "start_time": "2024-01-01T10:00:00Z", "end_time": "2024-01-01T10:05:00Z" }
Output: { "threats_detected": true, "incidents": [...], "analysis": "..." }
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.