• TEST 6: test of CAR at the ingress point of the network (*)
• TEST 7: impact of normal/exceed burst sizes on TCP throughput (Uni Twente) (*)
  additional test results by DANTE (same results as at Uni Twente)
  additional test results at Uni Utrecht (confirming results collected at the Uni. of Twente

• TEST 7: test of CAR and WFQ at the ingress of the network and UDP test traffic (*)
• TEST 8: test of CAR and WFQ at the ingress of the network and TCP test traffic (*)
• TEST 9: test of CAR with TCP rate limited traffic and TCP or UDP background traffic (*)
• TEST 9 b: like TEST 9 (CAR exceed action: set-prec-transmit 5), Uni Twente - inconsistency of results (*)
  additional test results at Uni Utrecht - inconsistency of results - different from the ones at INFN and Uni. of Twente (*)
• TEST 9 c: like TEST 9 (CAR exceed action: drop), Uni Twente (*)
• TEST 10: additional testing of CAR and WFQ as suggested by C.Filsfils (*)
• CAR + WFQ:

(*) IOS c7200-js-mz_120-5_0_2.T1, router platform: C7200

1. Baseline test: CPU utilization of a FastEth VIP2 with CAR as a function of the packet rate
2. Number of CAR entries (FastEth): influence on CPU utilization of the number of CAR entries
3. Match position:
   1. in a list of CAR commands
   2. within an ACL of several entries

• TEST INFN -- > CERN
  and trace of the TCP connection captured through snoop at the tx side (INFN)
• TEST CERN -- > INFN

• Precedence marking
• WFQ counters management

• Normal and exceed burst configuration under CAR
• CBWFQ configuration
• Test methodology for comparison of CAR and WFQ counters
• Cisco commands for queue monitoring