Research Topics

Hot Research Topics PDL

 

There are several fields which we are working on. Here you can find each Ph.D. student research work:

 

  • Marzieh MalekimajdCloud Computing Resource Management
  • Shirin BaghoolizadehProbabilistic Model Checking Game-Theoretic Verification Methods
  • Leila RashidiAnalysis and Improvement of Routing in Large Scale Delay Tolerant Networks
     
    • Large scale delay tolerant networks become more important as number of mobile devices increases every day. They have applications in smart city, cellular traffic offloading, etc. Dealing with state space explosion problem is an important challenge in analysis of routing protocols in large-scale heterogeneous DTNs. So, the main focus of this thesis is on proposing scalable Markovian framework/models for performance evaluation of routing protocols considering network heterogeneity.
       
  • Ehsan AtaieCloud Computing Performance Modeling and Evaluation
     
    • Cloud computing is an emerging commercial computing paradigm with its ease of access and diverse applicability. Modeling and evaluation of performance, dependability, and power consumption in cloud data centers are key concerns that should be investigated. The main focus of this thesis is on the simultaneous computation of performance, dependability, and power consumption metrics when different cloud management strategies and policies are considered. The approach will use mathematical tools like Markov chains, Petri nets, stochastic activity networks, and stochastic reward nets.
       
  • Maryam BagheriFormal Methods &Verification Stochastic Model Checking
  • Soroush KarimianBig Data Performance Modeling and Evaluation
     
    • In the era of Big Data which digital industry is facing the massive growth of data size and data dependent software, more and more businesses are moving to use new frameworks and paradigms capable of handling Big Data. The outstanding MapReduce (MR) paradigm and its implementation framework, Hadoop are among the most referred ones. Configuring a Hadoop cluster needs to set a lot of parameters, and great work has been done to optimize the configuration. One of the main difficulties in the way of Hadoop configuration optimization is the problem of estimating a simple MR Job execution time or the execution time of a more complex Tez workflow. In this regard, we propose analytical models based on SAN formalism to accurately model the execution of a MR or Tez job in the Hadoop framework governed with the YARN dynamic resource management layer or Spark framework. We validate the accuracy of the proposed models over industrial benchmarks running on Hadoop clusters with different configurations.
       

The alumni Ph.D. students also are contributing with PDL laboratory. You can find more information about their current research field in their websites and get in touch to see whether you can work with them or not. You may want to have a quick glance over their Ph.D. thesis abstract below:

 

  • Mohammad GharibKey Management for Large Scale Mobile Ad-hoc Networks
     
    • Mobile ad hoc networks have been attracted the attention of many researchers during last years. One of the major concerns faces such networks is the security issue. The root of this concern is the fact that the intermediate nodes have the responsibility of packet transportation and forwarding. The nodes inside the mobile ad hoc networks assumed to be trusty while they can read, change or drop the transported packets. Cryptography as the cornerstone of the security could be play an essential role in such networks. Any cryptosystem need some keys to be able to secure communications. In large scale mobile ad hoc networks storing the whole keys in all nodes is inefficient, if possible, due to the MANET's limitations in storage, energy and processing capability. In this thesis, considering MANET limitations and characteristics, two key management algorithms are proposed. The first one is aprobabilistic asymmetric key management based on key pre-distribution. We analytically prove that storing just a few keys in each node, keeps the network connected with very high probability. Such an idea faces many concerns and issues including routing problem, path length, the amount of traffic generated by the proposed algorithm and etc. In this thesis efficient solutions are proposed for the whole issues and concerns. The second proposed key management system is a fully distributed key management built based on the identity based cryptosystems, eliminatingthe need for storing keys. Both of proposed algorithms are analyzed analytically and compared with the recently proposed literature works, using simulation. Results show superiority of both of the proposed key management systems in comparison with the related works in both security strength and performance aspects.
       
  • Reza Entezari-MalekiPerformability Modeling and Analysis in Grid Computing
     
    • 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 serving 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 appropriately model and evaluate the combined performance and dependability of a resource, they encounter state space explosion problem whenever they get together to capture a real grid environment. To solve this problem, two approximate models are proposed using folding and fixed-point iteration methods to appropriately estimate the results of the monolithic model of a grid environment. The results obtained from numerical analysis of the proposed models and simulating the corresponding systems show that the approximate models can properly estimate the monolithic model.
      Afterwards, two scheduling approaches are presented to use the results gained from the previous step to schedule grid applications to grid resources. The first approach considers independent grid programs and uses curve fitting to find the service time of each resource to the set of independent programs, whereas the second approach applies the u-function technique to compute the probability mass function of service time of the entire grid to a workflow application consisting of many dependent programs. Finally, both approaches apply heuristic methods to solve the scheduling problem.