Techniques For Modeling Mobile Application Spreading

While the number of mobile devices and applications increases considerably, application spreading by using direct communication between mobile devices has not got too much attention so far. However, exploiting the advantages of modern communication paradi gms has even economic relevance. For instance, direct communication in ad hoc networks can be used to advertise and spread applications, which influences the number of purchases what application providers may want to estimate.In this paper, we present tw o main techniques for modeling application spreading in this environment. First, we model the spreading process using Closed Queuing Networks, assuming a single type user behavior. Then, to capture different user behaviors, we show how to model application spreading by Stochastic Petri Nets. We define a basic Petri net model that we analyze using a mean field based methodology. Unfortunately, introducing more realistic user behaviors makes the analytical handling of the Petri net too complex, or even imposs ible. We show an example of this case extending the basic Petri net model with an additional, more realistic user behavior, and investigate this extended model via simulations. Moreover, we compare the investigation results of the different models and poin t out their relations