This dataset has been used to generate part of the results included in the manuscript: "Cooperative Learning for End-to-End Delay Modeling in Multi-Domain Networks", submitted to IEEE Transactions on Network and Service Management (TNSM) at the time of publication of this dataset. The dataset contains synthetic end-to-end (e2e) delay and traffic measurements for multidomain paths established in a network consisting in three domains. Paths traverse either two or three domains. Besides e2e monitoring data, routing information of the paths and the models used to compute the delay introduced for each domain (i.e. intra-domain delay components of total e2e delay) are provided. Traffic and delay data were synthetically generated according to the CURSA-SQ methodology. For further details about it, please refer to the following reports: "The Logistic Queue Model", Tech. Rep. UPC-DAC-RR-GEN-2020-1 (Available online: https://people.ac.upc.edu/lvelasco/docs/research/UPC-DAC-RR-GEN-2020-1.pdf) and "Validation of the CURSA-SQ Methodology", Tech. Rep. UPC-DAC-RR-GEN-2020-2 (Available online: https://people.ac.upc.edu/lvelasco/docs/research/UPC-DAC-RR-GEN-2020-2.pdf)