Decisions made by the adaptation engine are currently
solely rule-based. Here, a self-learning approach is a promising
direction for future research, as events form a feedback-loop
between mobile devices and the game service. This loop can
be utilized to monitor the performance of the mobile peer-topeer
dissemination scheme itself and to adapt the respective
configuration accordingly. As shown during the evaluation,
merging of groups utilizing different dissemination protocols
can be further improved. To this end, the adaptation engine
could employ clustering algorithms to detect groups [1] that
are not limited to the AoI of a node. While this reduces additional
overhead at the central node, it can reduce the number
of dropped events due to locally mismatching dissemination
protocols. Studying the performance vs. cost trade-off of more
complex strategies at the adaptation engine is part of our future
work