The fast loss recovery mechanism of the AIDC

Simulation data of packet loss fast recovery mechanism in intelligent computing center: This dataset is based on the ns3 simulation tool to simulate three common data center RDMA congestion control algorithms, dcqcn, hpcc, and timely. In the simulation environment, packet loss scenarios are constructed and flow completion times are measured for each congestion algorithm to analyze their ability to handle packet loss events. dcqcn_gbn_output_m21.txt： Based on the dcqcn algorithm and gobackN algorithm, test adding one more flow. The first line after the diagonal line in the dataset is msg size, where min represents the minimum flow completion time, avg represents the average flow completion time, max represents the maximum flow completion time, and 99% fct represents the flow completion time at the 99% percentile. dcqcn_sr_output_m21.txt： Based on the DCQCN algorithm and SR algorithm, test adding one more flow. The first line after the diagonal line in the dataset is msg size, where min represents the minimum flow completion time, avg represents the average flow completion time, max represents the maximum flow completion time, and 99% FCT represents the flow completion time at the 99% percentile. hpcc_gbn_output_m21.txt： Based on the hpcc algorithm and gobackN algorithm, test adding one more flow. The first line after the diagonal line in the dataset is msg size, where min represents the minimum flow completion time, avg represents the average flow completion time, max represents the maximum flow completion time, and 99% fct represents the flow completion time at the 99% percentile. hpcc_sr_output_m21.txt： Based on the HPCC algorithm and SR algorithm, test adding one more flow. The first line after the diagonal line in the dataset is msg size, where min represents the minimum flow completion time, avg represents the average flow completion time, max represents the maximum flow completion time, and 99% fct represents the flow completion time at the 99% percentile. timely_gbn_output_m21.txt： Based on the Timely algorithm and GobackN algorithm, test adding one more flow. The first line after the diagonal line in the dataset is msg size, where min represents the minimum flow completion time, avg represents the average flow completion time, max represents the maximum flow completion time, and 99% fct represents the flow completion time at the 99% percentile. timely_sr_output_m21.txt： Based on the Timely algorithm and SR algorithm, test adding one more flow. The first line after the diagonal line in the dataset is msg size, where min represents the minimum flow completion time, avg represents the average flow completion time, max represents the maximum flow completion time, and 99% fct represents the flow completion time at the 99% percentile.