Empirical evaluation based on offensive and defensive experiments
Evans et al. 2011 Leeuwen et al. 2016
It can be used to test defense mechanism responding to different attacks.
It is limited to the number of samples. Thus difficulty in evaluating MTD effectiveness in wide range.
Empirical evaluation based on simulated experiments
Zhuang et al. 2012 Zheng et al. 2016
It can analyze MTD effectiveness in different application scenarios and compares in vertical.
It evaluates MTD effectiveness based on prior knowledge, which is not capable of assessing MTD to unknown attacks.
Abstract evaluation based on mathematical model
Clark et al. 2013 Han et al. 2014 Carroll et al. 2014
It has the advantage of universality, thus being able to compare different MTD mechanisms.
Since it is difficult to abstract complicated environment and tasks in parallel, it may cause the deviation between constructed model and the actual one.
Effectiveness evaluation based on mixed analysis
Okhravi et al. 2014 Zaffarano et al. 2015 Eskridge et al. 2015 Lei et al. 2016
It combines the advantages of the above analysis methods.
It needs to ensure the consistency of data process and transformation using different methods.