Autonomous formation flying (flocking) of UAVs requires coordination of efforts and solutions in multiple research fields (robotics, wireless sensor networks to name a few). The Airborne Embedded auTonomOUs Robust Network of Objects and Sensors (AETOURNOS) project is designed as a federative, fueling platform and global evaluation tool. It has spawned another collaborative project bridging the gap between Computer Science and Automation: 6PO (Cyber-Physical Systems and Cooperative Control of a Fleet of UAVs). We want to investigate new flocking strategies, mobility-aware communication protocols (and reciprocally: movements influenced by network QoS), and to ease the cooperation between experts, we need co-simulation solutions, and "universal" control interfaces. This article presents our first contribution to this grand scheme: a stand-alone simulation of an autonomous flocking strategy taking into account both the dynamics of the UAVs and their communications, which has allowed us to examine the relation between network unreliability and control computation. To better study the influence of the communication network, we present an architecture coupling a dedicated network simulator with the UAVs simulator, as well as with a real drone piloting interface. The lessons learned help us build a research and action plan for the years to come.