Research Interests

  • Network Science for Communications

  • Cloud Computing

  • Smart Grid, Smart Grid Communications

  • Network Resource Management

  • Traffic Engineering

  • Network Planning

  • Autonomic Networking

  • Optimization Theory and Applications

  • Mathematical Theory of Networks

  • Graph Theory and Network Analysis

  • Control Techniques for Data Networks

  • Network Algorithms

  • Green Communications

  • Wireless and Vehicular Networks

  • Next Generation Data Centers

  • VoIP

  • Bioinformatics
     

Thesis Summary

In my PhD thesis, I built a conceptual architecture for self-managing systems, with special emphasis on core networks where robustness to environmental changes are viewed as critical. The main objective of my thesis was to find a theoretical framework to build, control, and maintain robust communication networks. Motivated by ideas from Evolutionary Science, I tried to find a survival value for communication networks to quantify their sensitivity to the environmental changes (changes in topology, traffic, QoS parameters, and community of interest). I used random-walk betweenness (a metric from graph theory) of a link per unit of link weight to capture the effect of topology, load, and community of interest. I proved that this quantity is independent of the link location (therefore, it is a global measure on the graph), and it is a strictly convex function of link weights. I named this global value and used it as the survival value of a network. In order to have a robust network, we need to minimize Network Criticality.

As a part of my thesis, I developed a conceptual architecture for self-organizing management systems, AutoNet, to control the allocation of resources to the customers in order to maximize the robustness. The main idea was to consider a management system as an evolutionary process that evolves based on survivability and performance requirements. For my purposes, the evolution should be in the direction of decreasing Network Criticality, which in turn increases the network robustness. AutoNet consists of two autonomic loops, the slow loop to control the long-term evolution of robustness throughout the whole network, and the fast loop to account for short-term performance and robustness issues. AutoNet assigns separate Virtual Networks (VN) to different customers based on their specific Service Level Agreements (SLA), and the autonomic control loop tries to keep the customers' resources at the required level without service interruption. I used Network Criticality as the main control parameter and proposed appropriate algorithms for slow and fast loops of AutoNet, to guide the network evolution in real time. For the fast loop, I developed some flow assignment algorithms to engineer the traffic according to the customer demands, and for the slow loop, I used Semi-Definite Programming (SDP) approach to solve the convex optimization problem for Network Criticality in order to do network (re)planning which is the main job of the slow loop.

 

Teaching Experience

Instructor

1. University of Ontario Institute of Technology (UOIT), Oshawa, Canada, Winter 2011,  Course: "Modern Control Systems (ENGR 3100U)".

2. University of Ontario Institute of Technology (UOIT), Oshawa, Canada, Fall 2010,  Course: "Digital Communications (ENGR 4130U)".

3.Azad University of Tehran, Department of Computer Science (CS), Tehran, Iran, 1996-1999.

4. Azad University of Saveh, Department of Electrical \& Computer Engineering (ECE), Tehran, Iran, 1998-2001.