1. 1 Authors: Scott Poretsky, Reef Point Systems Shankar Rao, Qwest Communications 64th IETF Meeting – Vancouver Accelerated Stress Benchmarking draft-ietf-bmwg-acc-bench-term-07.txt draft-ietf-bmwg-acc-bench-meth-04.txt draft-ietf-bmwg-acc-bench-meth-ebgp-00.txt draft-ietf-bmwg-acc-bench-meth-opsec-00.txt

2. 2 BMWG Stress Work Item BMWG is in the Operations and Management Area and writes recommendations for benchmarking the performance characteristics of network devices Stress Test provides rigorous benchmarking for deployment readiness: –Verifies operation in deployed environment –Ensures System and Software stability under failure conditions –Validates correct routing under stress –Ensures device manageability under stress –Identifies difficult to detect software bugs –Increases ability to survive DoS attacks Stress test coverage has direct relationship to Operational Security

3. 3 BMWG Accelerated Stress Document Structure Terminology EBGP Peering Methodology … General Methodology Operational Security Methodology –Additional technology specific methodologies can be added

4. 4 Stress Test Configuration Instability Conditions BGP OSPF Multicast IPsec Data Plane Module Management Plane Module Security Plane Module...... Router/Switch DUT Control Plane Module Startup and Instability Conditions Controlled by Tester or System of Testers Tests often need to be automated to achieve the repeated nature of stress conditions BGP OSPF IPsec Telnet/FTP Traffic Forwarding Events (Traps, Logs) Router/Switch DUT Router/Switch DUT Router/Switch DUT Router/Switch DUT Router/Switch DUT Router/Switch DUT Stress Test VERSUS Isolated Feature Test Startup Conditions

5. 5 Stress Test Methodology 1. Startup Phase 2. Instability Phase 3. Recovery Phase Apply Startup Conditions BENCHMARKS: Stable Aggregate Forwarding Rate Degraded Latency Stable Session Count Achieve Configuration Set Apply Instability Conditions BENCHMARKS: Unstable Aggregate Forwarding Rate Degraded Aggregate Forwarding Rate Average Degraded Aggregate Forwarding Rate Unstable Latency Unstable Uncontrolled Sessions Lost Remove Instability Conditions BENCHMARKS: Recovered Aggregate Forwarding Rate Recovered Uncontrolled Sessions Lost Recovery Time Variability Benchmarks

6. 6 DEVICE #1 1. Configuration Sets achieved 2. Startup Phase Benchmarks Stable Aggregate forwarding Rate = 4Gbps Stable Latency = 110 usec Stable Session Count = 30 BGP Peers 28 OSPF Adjacencies 16K IPsec Tunnels 3. Apply Instability Conditions 4. Instability Phase Benchmarks* Unstable Aggregate Forwarding Rate = 3.5Gbps Degraded Aggregate Forwarding Rate = 0.5Gbps Unstable Latency = 110usec Unstable Uncontrolled Sessions Lost = 126 *These are averages. It is recommended to record these values at 1 second interval 5. Stop applying Instability Conditions after X hours (24 for this test) 6. Recover Phase Benchmarks Recovery Time = 22 seconds Recovered Aggregate Forwarding Rate = 4Gbps Recovered Latency = 110usec Recovered Uncontrolled Sessions Lost = 0 Example Stress Test Benchmarks DEVICE #2 1. Configuration Sets achieved 2. Startup Phase Benchmarks Stable Aggregate forwarding Rate = 4Gbps Stable Latency = 150 usec Stable Session Count = 30 BGP Peers 28 OSPF Adjacencies 16K IPsec Tunnels 3. Apply Instability Conditions 4. Instability Phase Benchmarks* Unstable Aggregate Forwarding Rate=3.3Gbps Degraded Aggregate Forwarding Rate= 0.7Gbps Unstable Latency = 170usec Unstable Uncontrolled Sessions Lost = 4000 *These are averages. It is recommended to record these values at 1 second interval 5. Stop applying Instability Conditions after X hours (24 for this test) 6. Recover Phase Benchmarks Recovery Time= Infinite Recovered Aggregate Forwarding Rate = 3.9Gbps Recovered Latency = 150usec Recovered Uncontrolled Sessions Lost = 97 Iteration 1 – 1 st Configuration Set scaled larger but Device #2 crashed at 20 hours Iteration 3 - 3 rd Configuration Set scaled lower to obtain a Recovery Time for Device #2

7. 7 Next Steps Is Terminology and Methodology ready for WGLC? Incorporate mailing list comments from BMWG and OpSec Identify and Add more test cases to EBGP Peering and Operational Security Methodologies Suggestions posted on Mailing List for new Methodologies: MPLS-TE network specific test cases LDP over RSVP-TE specific test cases

8. 8 Backup

9. 9 EBGP Peering Methodology –draft-ietf-bmwg-acc-bench-meth-ebgp-00.txt, Methodology for Benchmarking Accelerated Stress with Operational EBGP Instabilities Operational Security Methodology –draft-ietf-bmwg-acc-bench-meth-opsec-00.txt, Methodology for Benchmarking Accelerated Stress with Operational Security EBGP Peering Stress Test Cases 4.1 Failed Primary EBGP Peer 4.2 Establish New EBGP Peer 4.3 BGP Route Explosion 4.4 BGP Policy Configuration 4.5 Persistent BGP Flapping 4.6 BGP Route Flap Dampening 4.7 Nested Convergence Events Operational Security Stress Test Cases 4.1 Restart Under Load 4.2 Destination Control Processor 4.3 Destination Control Processor with Rate-Limiting 4.4 Destination Interfaces 4.5 DoS Attack Current Stress Test Cases