Computer networking changed faster in the years leading up to 2007 than in perhaps any comparable stretch before it. Peer-to-peer file sharing, large-scale wireless sensor deployments, and IP telephony all moved from research curiosities to services used by hundreds of millions of people within a single decade. Each of those applications forced new protocol layers into existence, and cross-layer optimization — tuning a wireless MAC around what the transport layer needed, for instance — started to blur the clean, textbook separation between layers that networking courses had taught for twenty years.
That complexity is what pushed the field toward experimentation testbeds and simulation tools in the first place. Analytical models are elegant, but they struggle with systems whose behavior depends on dozens of interacting, non-steady-state processes. Simulation and the coupling of simulation with real testbeds became, in the co-chairs’ view, the more honest way to evaluate a protocol whose performance genuinely could not be derived on paper.
The first edition
For NSTools’07, the inaugural edition, the workshop received 40 full-paper submissions from authors on every continent. The program committee accepted 10, organized into three sessions that mapped onto the workshop’s three founding themes: the evaluation of simulation accuracy, domain-specific modeling choices, and the development of ad hoc tools built to fill gaps in existing simulators. Prof. Rajive Bagrodia opened that first edition with a keynote on WHYNET, a framework for in-situ evaluation of heterogeneous mobile wireless systems.
Why the same themes still hold in 2026
Two decades later, the specific tools have changed, but the underlying problem has not. The ns-3 project now ships a new release roughly two to three times a year, each one adding validated models for technologies — low-earth-orbit satellite mobility, updated Wi-Fi power-save behavior — that simply didn’t exist when ns-3 began. That release cadence is only trustworthy because of exactly the kind of validation-focused research NSTools was built to showcase: without a community checking simulated behavior against real systems, a simulator’s outputs are just numbers, not evidence.
The same is true of testbeds. Independent research comparing simulators, emulators, and physical testbeds continues to find the same trade-off the co-chairs described in 2007: simulators and emulators offer control and reproducibility at the cost of realism, while testbeds offer realism at the cost of scale and convenience. Choosing correctly, for a given research question, is still a skill — and still worth a dedicated technical program.
Editorial note — who wrote this, and why
This page preserves the original 2007 co-chairs’ message, lightly edited for length, alongside added context connecting it to current simulation tools. It is maintained by the NSTools Program Committee as a historical record of the workshop’s founding rationale; see the 2026 technical program for this year’s active research content.
Acknowledgements
The founding co-chairs thanked the members of the first Technical Program Committee for completing a large volume of reviews on short notice, and thanked Tania Jimenez and Sara Alouf for their guidance through each step of organizing the first edition. They also thanked the Valuetools steering committee and the workshop’s sponsors. That structure of volunteer reviewing and technical co-sponsorship, now shared with ICST and CREATE-NET, is the same one that keeps NSTools running each year.