Abstract :
Configuring large-scale network devices that are built manually takes a long time, this can be overcome by network automation. With this network automation configuration can be done automatically, so that routing configuration can be done in a short time and reduce human errors. The routing configuration uses the Python programming language that supports the programming language with the types of libraries used in the research, namely Paramiko and Napalm for comparison. Paramiko can be used as an SSH Server or Client, while Napalm supports multivendors that will make comparisons using the BGP protocol used between AS so that it can reach even remote areas. In this study, the results of the automatic BGP configuration time using Paramiko were faster with an average value of 5.365 seconds compared to using Napalm. In addition, the use of Paramiko in performing BGP convergence between clients obtained a faster value with an average value of 182,300 seconds compared to Napalm. The throughput value in automation using Paramiko has a better average value of 10,174 Mbps compared to Napalm, but if the throughput is done by providing data services automatically or manually, the throughput value obtained is the same as an average of 2,945 Mbps. The delay value performed using automation with the Paramiko library has a faster value than the Napalm library with an average value of 1.475 ms, in contrast to the delay that is configured automatically or manually followed by data services to obtain the same average value, namely 55,336 ms. The results of measurements based on TIPHON standardization, the delay value is categorized as "very good" with a value of