The experiments on WSAN testbed were conducted in only 3 environments: clean, noisy and stress testing. More environments with various condition of interference should be used in the experiment to more convincingly show the difference between graph routing and source routing.
The paper said regular flows should be delivered on a best-effort basis when an emergency occurs. But it didn’t describe exactly how this best-effort is implemented. It lacks theory analysis for the different methods of scheduling emergency packets such as slot stealing and event-based emergency.
The experimentation and results in this paper are only for WirelessHART. It is better to conduct experiments using other WSAN standards as well and show whether the methods and algorithms are equally applicable to other standards.
There are other open issues I think worth considering. For example, whether the network can automatically choose the best routing algorithm according to the environments. Or whether there are better routing algorithm than the source routing and graph routing in the WirelessHART standard.
It is impressive that the paper authors applied knowledge in other domains to the problem of WSANs. For example, state-of-the-art response time analysis for multiprocessor scheduling is used for the priority scheduling in WirelessHART net- works. The authors made use of the Simulink from control system community and TOSSIM from sensor network community to conduct simulations for wireless cyber-physical simulation. Since problems involved in WSAN come from many different domains, such kind of Interdisciplinary combination is very valuable.