OVERVIEW

• General information
• Programme

  EXPERIMENTAL ACTIVITIES AND RESULTS

• Shaping: Experimental results
• Scheduling: Deficit Round Robin and Priority Queuing
• Scheduling: Weighted Round Robin, policing and marking (Catalyst 6500)
• TCP performance: optimization through ECN

  SERVICES

• Less-than-best-effort

• Reading

General information

• Mailing list archive

Work programme

Hardware requirements

Experimental activities and results

1. Experimental results on shaping

2. Deficit Round Robin and Priority Queuing
   • MDRR performance measurement results
   • Example of queueing architecture of router based on DRR: Cisco GSR and Engine 2 interfaces (pdf)
   • Deficit Round Robin, old testbed information about Plage test network

3. QoS testing on Catalyst 6500

   Policing:

   • Test 1: Dual Token Bucket policing on Catalyst 6500 for TCP traffic, Experimental results
   • Test 2: Analysis of a TCP stream profile generated between two GigaEthernet hosts; Results
   • Test 3: Policing applied to UDP traffic; Results

     Weighted Round Robin:

   • Test 4: tuning of queue weight configuration with UDP traffic; Results
   • Test 5: test of traffic isolation with variable number of WRR classes; Results
   • Test 6: effect of threshold parameter in per sub-queue traffic isolation; Results
   • Test 7: dependency of TCP throughput on parameter queue-limit; Results

4. TCP performance optimization
   • Experimental results

Services

• Less-than-best-effort

Suggested reading

1. Fast Switched Backplane for a Gigabit Switched Router, Nick McKeown, Stanford University (pdf)
2. Cisco 12000 Series Internet Routers Technical Documents, (html)
3. The evolution of high-end router architecture - Basic Scalability and Performance Considerations for Evaluating Large-Scale Router Designs, (html)