Probabilistic model checking for the Grid and on the Grid
Marta Kwiatkowska
School of Computer Science, University of Birmingham
www.cs.bham.ac.uk/~mzk
The increasing dependence of society on computer systems, from desktops through
handheld devices to grid systems, deployed in business-critical domains requires
assurances of software correctness and performability. Formal verification techniques
such as model checking have proved very successful in this area. Model checking
involves the construction of a model of a real-life system, formal specification of
system’s properties, and then an automated exhaustive analysis of the model. There
are many examples of real-life systems for which an accurate analysis must also take
into account stochastic aspects of the system’s behaviour, for example randomisation
used as a symmetry breaker, uncertainty in user behaviour, or unknown
communication delays. Probabilistic model checking [2,3], and in particular the
probabilistic model checker PRISM [4,1], have proved very successful in discovering
design flaws and performance characteristics of a range of real-world examples [5].
In common with conventional model checking techniques, probabilistic model
checking suffers from state-space explosion and therefore extending the range of
systems that can be efficiently modelled and analysed is a major challenge.
The aim of this talk is two-fold. Firstly, we will demonstrate the usefulness of
probabilistic model checking for Grid systems by giving examples of performability
studies of computer clusters. Secondly, we demonstrate how probabilistic model
checking can be enhanced through the application of parallel computing [6] and, in
particular, Grid computing [7]. Finally, we will summarise the main challenges
facing this area.
1.
2.
3.
4.
5.
6.
7.
PRISM website www/cs.bham.ac.uk/~dxp/prism/.
J. Rutten, M. Kwiatkowska, G. Norman and D. Parker, Mathematical Techniques for
Analyzing Concurrent and Probabilistic Systems, CRM Monograph Series, vol. 23, AMS,
2004.
M. Kwiatkowska, Model Checking for Probability and Time: From Theory to Practice. Invited
paper, In Proc. LICS'03, p 351-360, IEEE Computer Society Press, 2003.
M. Kwiatkowska, G. Norman and D. Parker, PRISM 2.0: A Tool for Probabilistic Model
Checking. In Proc. QEST'04, pages 322-323, IEEE Computer Society Press, 2004.
M. Kwiatkowska, G. Norman and D. Parker, Probabilistic model checking in practice: Case
studies with PRISM. ACM Performance Evaluation Review on Performance and Verification,
32(4), p 16-21, 2005.
Y. Zhang, D. Parker and M. Kwiatkowska, A Wavefront Parallelisation of CTMC Solution
using MTBDDs. In Proc. Dependable Systems and Networks (DSN'05). To appear.
Y. Zhang, D. Parker and M. Kwiatkowska, Grid-enabled Probabilistic Model Checking with
PRISM. In Proc. All Hands Meeting (AHM'05). To appear.