Simplicafication of resource constraints in generalized stochastic petri net-based software performance modelsGeneralized Stochastic Petri Net 기반 소프트웨어 성능 모델에서의 자원제약 표현의 간소화

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 550
  • Download : 0
Generalized Stochastic Petri Nets (GSPN) has gained much attention in software performance modeling because it provides not only performance quantities, but also qualitative analysis power by means of the underlying Petri nets. However, GSPN-based performance modeling has two well-known drawbacks. The one is its computation difficulty for models of large systems. The other is the inefficiency in model constructions. For the latter drawback, it is related to model size and inconvenient notations. This thesis is to tackle the inefficiency. As a software system grows, the size of its corresponding model also grows. It may be true for any graphical modeling language. In the point of view of performance modeling, GSPN is weaker than other modeling language; e.g. Queueing Networks. The major reason of the relative weakness is because the graphical notation of GSPN is too general; coincidence of its generality. In the software performance modeling, however, a well defined modeling pattern exists; modeling of resource constraints. Furthermore, the resource constraints cannot be omitted in software performance models. Therefore the size of software performance models can be reduced by simplifying the well known pattern as a set of convenient notations. By using the simplified notation, a software performance model can be reduced in its size while conveying the same meaning more intuitively. The reduction may also lead to less effort in creating a software performance models. On the other hand, the performance quantities of the model must not be changed from the original model. Thus, a systematic unfolding mechanism or process has to be provided to obtain the corresponding GSPN models from the simplified models. The systematic unfolding process is about to be implemented on GSPN tools. Then, modelers can construct a software performance model with convenient notations and retrieve performance quantities by means of a capable tool. Therefore, this thesis provides a ...
Advisors
Baik, Jong-moonresearcher백종문researcher
Description
한국과학기술원 : 소프트웨어공학 프로그램,
Publisher
한국과학기술원
Issue Date
2010
Identifier
455181/325007  / 020084311
Language
eng
Description

학위논문(석사) - 한국과학기술원 : 소프트웨어공학 프로그램, 2010.08, [ ix, 65 p. ]

Keywords

Generalized Stochastic Petri Nets; Software Performance Model; Performance Modeling; 성능 모델링; Generalized Stochastic Petri Nets; 소프트웨어 성능 모델

URI
http://hdl.handle.net/10203/40127
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=455181&flag=dissertation
Appears in Collection
ICE-Theses_Master(석사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0