Enhancing Global Network Monitoring with Magnifier
summary
Magnifier infers ingress and egress prefixes from sparse packet samples then validates those inferences by deploying negative mirroring rules on routers to monitor where traffic should not appear. The system runs on commercial routers, uses Kaida traces for evaluation, and in simulation reaches up to 77% validated coverage (static assignments) versus ~15% for random; lab tests add up to 500 mirroring rules per device, generate at most 1.4% mirror traffic, and surface misinferences, failures and some DDoS events.
AI-written summary · about
chapters (10)
- 0:00Network traffic monitoring challenges
- 1:30Current traffic monitoring limits
- 3:36Magnifier's combined approach
- 5:23Inferring traffic ingress points
- 6:47Validating with traffic mirroring
- 7:50Scaling Magnifier's deployment
- 9:15Performance in simulations
- 12:24Real-world testing with Cisco
- 13:49Comparison with other systems
- 14:10Conclusion and future
related talks
25mStroboscope: Declarative Network Monitoring on a Budget
21mMirrorNet: High-fidelity and Scalable Network Emulation for Software-defined WAN
46mLessons learned from monitoring a large production network
51mAdvanced approaches to traffic inspection & network diagnosis w/
32mActive Fault Finding in Networks
16mResolving Packets from Counters: Enabling Multi-scale Network Traffic Super Resolution...
49mVisibility Without Borders: NETSCOUT to help migrate apps to
16mClosed-loop Network Performance Monitoring and Diagnosis with SpiderMon