Existing modeling approaches divide according to various criteria. We can find the related works grouped into publications based on stochastic models such as Queueing Net (classical product-form, extended or layered) and stochastic models such as Petri Nets. In the first group, there are some opened and closed queueing models. In [

7] the author presents the model application of server cluster and database server. In the paper, [

8] authors propose the model software systems using classical queuing theory and statistical response time models in parallel. This approach allows users to tailor the model for each analysis run, based on the performed adaptations and the requested performance metrics. In [

9] the author prepared the multi-tiered application model by using Mean Value Analysis (MVA) and compared performance with real performance values. Recent publications [

2,

10] use QPN for performance modeling of Web structure and compare with the production system performance. To manage the routing of packages in Web structures, we can apply elements of the control theory. In turn, in the second group, there are different variants of the Petri Nets [

10,

11]. In [

12] for some modes of emergency actions, Petri Nets models are proposed. The Apache Storm model was presented in [

13]. The author used Generalized Stochastic Petri Nets for performance checking. Proposed in [

14] techniques of analysis could help in the context of Web systems performance, that nowadays confront the era of Big Data. Various papers propose the arrival of stream modeling and its analysis. According to the Web system modeling method introduced in [

10] the modeled system (Platform as a Service designed to host Web applications) retains a structure similar to an open queueing system. Analysis of the distribution of data packages within the system makes it possible to find out some potential bottlenecks [

6]. To efficiently carry out the randomly arriving data [

15], the number of overlapping (concurrent) request/response sessions utilized by the sending node, is higher than in typical client–server communications.

Furthermore, there are some mathematical models for predicting the performance of the Web system. Almost in all cases, the authors conducted experiments [

5,

16]. This is the second way for performance engineering. The use of simulations models and experiments allows verifying the proposed models. Moreover, it greatly influences the validity of the systems’ development. In the article [

17] they examined request distribution strategies used in one-layer and two-layer architectures of Web systems. In paper [

18] they discuss current software architectural styles, patterns, and development platforms based on client-side and server-side technologies. Some authors in [

19] use Production Trees, the new modeling methodology for dealing with availability issues of production systems.