Sharif Digital Repository

Businesses are dependent on mining of their Big Data more than ever and configuring clusters and frameworks to reach the best performance is still one of the challenges. An accurate performance prediction of the Big Data application helps reduce costs and SLA-violations with better tuning of the configuration parameters. Among the Big Data frameworks, Hadoop, Tez, and Apache Spark are the widely used and popular ones, with the MapReduce and graph-based workflows, usually running on top of the YARN cluster. While a great number of attempts have been made to predict the execution time of Big Data applications, to the best of our knowledge, none of them considered multiple simultaneous YARN... 

Expansion planning of distribution network assets is of importance to reliably supply the growing demand. Growing penetration of distributed energy resources (DERs) including renewable based distributed generators (DGs), conventional DGs, energy storage systems, and demand response (DR) resources in the network as well as their operational advantages make it necessary to address the investment planning of these resources in addition to the network assets. In order to effectively utilize the DERs and to solve the challenges arisen by their increasing penetration, active distribution networks (ADNs) are introduced which offer active network management capability. Passing through traditional... 

In recent years, we have seen the proliferation of real-time and time-dependent systems. They have increased both in number and in complexity. Due to their increasing complexity, we need to model them and become assured of their correctness. Several formalisms have been devised for time-dependent systems. They include timed automaton and timed extensions of Petri nets. Our focus is on timed extension of activity networks and its model checking. Activity networksare powerful formalisms which allow designers to model the system in a natural, small and conceivable manner. This way, timed extension of activity networks can greatly help modelers. In this thesis, we propose a method to specify the... 

Availability of the service delivered by cloud providers is one of the most important QoS factors of the service level agreements between providers and customers. Since current Infrastructure-as-a-Service providers use virtualization technology to manage data centers,virtual data centers (VDCs) have become a popular infrastructure for cloud computing.Evaluation of VDCs gives cloud providers the oportunnity to improve their service by investigating the effect of different parameters on their service.As a result, they can allocate appropriate amount of resources, and apply better resource management policies.In this study, in order to evaluate different aspects of VDCs, including service... 

Blackhole and grayhole attacks have been become two of the major security concerns in mobile ad hoc networks (MANET). To achieve security in MANETs, a lot of mechanisms had been proposed by now. Using intrusion detection systems(IDSs) is one of the important mechanism to reach this goal. Thus, a well-known IDS was chosen and analyzed in this thesis. Furthermore, a collaborative bayesian filter approach for this intrusion detection system was proposed to enhance its performance. Then the performance of this approach was considered. This intrusion detection system was analyzed using stochastic modeling like continuous time markov chain(CTMC), stochastic reward net(SRN) and stochastic... 

Many studies has been focused on extracting structural and behavioral properties of complex networks in recent decade. Online Social Networks (OSNs) are one example of complex networks. Nowadays with rapid growth of OSNs such as Facebook and Twitter, the study of OSNs has become an interesting research area. Many of recent OSN studies studied friendship networks. Friendship network is a binary unweighted network, and all of its links have the same importance. But, in reality not only all friendship links are not representative of social interactions, but also the social links have a variety of intimacy, intensity, and etc. So, all links should not be considered equal. Recently researchers... 

Some systems exhibit probabilistic and nondeterministic behavior. This behavior arises, for example, due to the use of randomized algorithms or presence of the unreliable components. Probabilistic Activity Network (PAN) is a flexible model for describing such systems. Every PAN is convertible to a Markov decision process. Up to now, there is no model checker which can operate directly on this model. In addition state space explosion arises as an important problem in model checking of real systems. We have developed a software tool called SPTMAN, for model checking of PAN. In this dissertation we are going to present a way for applying partial order reduction on PAN and then we will introduce... 

Modeling and performance evaluation of the real time distributed systems is a significant problem. Stochastic activity network is one of the high level models used for this purpose. This network is an extension of generalized stochastic Petri net which is more powerful and flexible than other Petri net extensions. Checking the satisfiability of properties such as performance, dependability and user's expected properties is considerable in these models. Thus, using the approaches and generating modeling and model checking tools with the mentioned purpose is attractive. A model checker automatically checks the correctness of the system behaviors as properties against the model by getting the... 

In this thesis, three different mathematical models named Markov Reward Model (MRM), Stochastic Reward Net (SRN) and Stochastic Activity Network (SAN) are used to model and evaluate the performability of grid computing environments consisting of many grid resources. The proposed models consider the arriving and servicing process of grid tasks inside a resource together with the failure-repair behavior of processors of the resource. Since the proposed MRM cannot be extended to model a grid environment with some realistic assumptions, we switch to use SRNs in modeling a single grid resource with more number of processors. Although the proposed SRN models for a single grid resource can... 

Verification of timed systems is concerned by many researchers in the last decades because of using these systems in critical applications. So, several models such as Timed Automat (TA) and Time Petri Nets (TPN) have been proposed for modeling these systems and variety of logics and abstraction methods are defined. One of the significant problems in timed system is cost-optimal reachability problem where, given an initial state and a target state, the task is to comput a path with infimum cost from initial state to the target state. Priced Timed Automat (PTA) introduce for solving cost-optimal reachability problems such as optimal scheduling and planing in timed systems. PTA is defined as an...