October 19, 2026 · Pisa, Italy
Technical program
A single-track day: one keynote and three paper sessions covering monitoring frameworks, channel modeling, factorial analysis of simulation output, queueing theory, and peer-to-peer storage.
A single-track day: one keynote and three paper sessions covering monitoring frameworks, channel modeling, factorial analysis of simulation output, queueing theory, and peer-to-peer storage.
Why this schedule is structured the way it is. Each session groups papers by one of NSTools’ three founding themes rather than by tool or protocol, so that a paper on Wi-Fi channel modeling sits next to a paper on a switch-module implementation: both are, at their core, about whether a simulated component behaves like the real one. That grouping is intentional — it is what lets attendees compare validation methodology across very different sub-fields in the same room.
Session times are in Central European Time (CET). A printable version of this schedule is available on request from the workshop organizers.
Random number generation is not a peripheral detail in network simulation — every stochastic arrival process, every simulated packet loss event, ultimately depends on the statistical quality of the generator underneath it. Weak generators can silently bias simulation results in ways that are difficult to detect after the fact, which is why this keynote opens the program rather than appearing as a regular session paper.
10:30 AM – 11:00 AM: Coffee break
All three papers in this session extend the same open-source codebase, ns-3, which the ns-3 project describes as a discrete-event simulator maintained by a worldwide volunteer community under the GPLv2 license. That a monitoring framework, a PHY-layer channel model, and a switch module can all plug into the same simulation core is a direct consequence of ns-3’s modular, library-based architecture rather than a single monolithic GUI tool.
12:30 PM – 2:00 PM: Lunch break
Queueing models like M/M/1 look deceptively simple on a whiteboard, but reproducing their steady-state behavior inside a discrete-event simulator raises real implementation questions — how events are timestamped, how ties are broken, and how random draws are seeded. The ns-3 project’s own reference manual documents exactly this kind of subtlety: concurrent events at the same simulated timestamp are resolved in FIFO order by a monotonically increasing event ID, a design choice that directly affects reproducibility.
3:30 PM – 4:00 PM: Coffee break
Passive inference — estimating a flow’s congestion state without instrumenting the endpoints — is exactly the kind of question that is hard to answer with a physical testbed alone, since ground truth about the sender’s actual congestion window is rarely observable externally in a live network. Simulation provides that ground truth for free, which is why this closing session pairs naturally with the workshop’s broader theme of matching evaluation method to research question.
Session content, times, and author affiliations reflect the program as accepted by the NSTools 2026 Program Committee following peer review. If you are an accepted author and notice an error in your paper’s title or affiliation listing, contact the workshop organizers using the address in the footer so it can be corrected before the printed proceedings are finalized.