This repository contains evaluation instances used in the paper "TROPIC: Traffic-Engineering-Oriented Planning of IP Core Networks" (to appear at IFIP Networking 2025).
If you find this data useful in your research, please cite the following BibTeX entry:
@inproceedings{richardt-tropic,
author = {Leon Richardt and
Alexander Brundiers and
Timmy Sch{\"{u}}ller and
Nils Aschenbruck},
title = {TROPIC: Traffic-Engineering-Oriented Planning of IP Core Networks},
booktitle = {{IFIP} Networking Conference, {IFIP} Networking 2025, Limassol,
Cyprus, May 26-29, 2025},
publisher = {{IEEE}},
year = {2025},
note = {(to appear)},
}The evaluation dataset consists of 12 instances. Each instance consists of the following:
- A network topology (
.graphfile extension) - A traffic projection (
.demandsfile extension)- Additional information about the projection is contained in the
.projected-demands-info.jsonfiles
- Additional information about the projection is contained in the
- Two sets of candidate links (
.jsonfile extension)
The topology and traffic files follow the REPETITA file format.
The network topologies are based on their original versions in the Internet Topology Zoo, as contained in the REPETITA dataset. The topologies were pre-filtered for those we deem most representative for the studied use case. Additionally, the topologies have been slightly modified to make them more suitable to capacity planning.
Refer to Section V-A of the TROPIC paper for details.
The traffic data is based on the original traffic data contained in the REPETITA dataset, which has been generated by the Random Gravity Model. The demand file in the REPETITA dataset is interpreted as "the current" traffic demands. A suitably parameterized log-normal growth model is then applied to each node's ingress and egress volumes before recomputing the traffic matrix following the Gravity Model. The resulting traffic matrix represents the traffic projection for the planning horizon.
Refer to Section V-B of the TROPIC paper for details.
Candidate link sets are represented in a custom JSON format: Each candidate link is identified by the indices of its tail and head nodes in the .graph file.
For each topology, we provide candidate links chosen according to the only existing and DSPS strategies.
Refer to Section V-C of the TROPIC paper for details.