Ph.D. students:

Mohammad Erfan Mazahri

Mohammad Mazaheri
(Ph.D.)

Network Security

Khalaji

Behzad Khalaji
(Ph.D.)

Software Security

Vahid Moghiss
(Ph.D.)

Network Security

Masoud Ganjkhani
(Ph.D.)

Network Security

Mayabi

Maryam Mayabi
(Ph.D.)

Cloud Computing

MS.c. Students:

Fatemeh Abbasi
(M.Sc.)

A Stateful Algorithm for Optimizing FaaS Placement and Scheduling

Edge computing has rapidly gained prominence as a key technology in information and communication systems, capturing the interest of both researchers and industry experts. A major challenge in such dynamic environments lies in managing the number and distribution of active function instances across resources, as this has a direct impact on both resource utilization and latency. This issue becomes particularly critical in edge computing, where high volumes of user requests lead to constant variations in the demand for function instances during operation. Moreover, the resource constraints of edge environments and the necessity to minimize delays in data processing highlight the importance of efficient strategies for function placement and scheduling. This study focuses on tackling these challenges by developing effective approaches for optimizing function placement and scheduling in edge computing systems. on Sample Description
Sajad Aqanasiri

Sajad Aqanasiri
(M.Sc.)

Vulnerability detection in smart contracts

Smart contracts play a vital role in blockchain ecosystems but are frequently exposed to vulnerabilities such as reentrancy attacks, overflow errors, and mismanagement of cryptographic keys. These vulnerabilities can lead to severe financial losses and undermine trust in decentralized systems. To address these risks, this research presents a framework leveraging fault modeling and machine learning to simulate real-world attack scenarios and enhance the detection of security flaws. By bridging gaps in traditional approaches, the proposed solution aims to strengthen smart contract security and ensure robust blockchain operations.
Alireza Safafard

Alireza Safafard
(M.Sc.)

Smart Contract between DataCenters

A federated environment consists of multiple autonomous service providers operating within a unified framework while maintaining their brand identities. This model reduces costs, enhances resource efficiency, and lowers carbon emissions. However, challenges include brand dilution, customer lock-in, restricted competition, and opaque profit distribution. To address these issues, this research proposes a blockchain-based federated model with smart contracts, optimizing VM placement, improving customer autonomy, ensuring transparent provider evaluation, and maintaining provider independence.

Mohammad Mahdi Teymourian
(M.Sc.)

Autoscaling of Serverless Functions

Function as a Service (FaaS) is a serverless computing model that allows developers to deploy small, independent functions that automatically scale in response to demand without the need to manage servers or infrastructure. In this research, we propose a solution to proactively autoscale FaaS to enhance performance and conserve resources.

Hamidreza Bagheri
(M.Sc.)

Proactive Placement of IoT Applications in Mobile Fog Environments

Fog computing was introduced to address the limitations of traditional cloud computing for applications requiring low latency, high bandwidth, and real-time processing, such as microservice-based IoT applications. However, most existing solutions in fog environments are reactive, adjusting only after network or performance issues arise. This becomes especially problematic when fog nodes are mobile, such as in vehicular networks or drone-based systems, where changing topologies lead to increased communication delays and degraded service performance. To address these challenges, we propose a proactive placement solution to reduce delays and minimize service migrations.

Fateme Esmati
(M.Sc.)

Adaptive Service Function Chain Placement with AI-Driven Bottleneck Identification and Cloning for Delay Mitigation

One of the key challenges in network function virtualization is the optimal placement of Service Function Chains (SFCs), which can be improved by executing functions in parallel and identifying potential delays. The aim of this research is to dynamically detect bottleneck functions within delayed chains and mitigate these delays through the application of function cloning.

Mostafa Jafari
(M.Sc.)

Adversarial Attacks on Android Malware Detection: Exploring Vulnerabilities and Enhancing Robustness

Current state-of-the-art Android malware detection systems rely on machine learning models. Despite their superior performance, these models are susceptible to adversarial attacks. In this research, we explore the vulnerability of Android malware detection systems to such attacks. We investigate techniques used by attackers to evade detection and propose strategies to enhance the robustness of detection mechanisms.

Alireza Yousefi
(M.Sc.)

Pattern-Based Optimization of Initial Service Function Chain Placement

One of the important issues in cloud computing is the placement of the service function chains, which can be optimized by using some patterns. The purpose of this research is to identify these patterns dynamically and to examine their impact.

Reza Ghazinour
(M.Sc.)

Improving the security of industrial control systems using internet search engine data

The security of industrial control systems is of great importance. In this research, firstly, we examine the scanning characteristics of the Shodan tool in terms of duration, time interval between scanning repetitions, scanned ports, scanned protocols, scanners' geographical area, and finally using Machine learning techniques, we try to recognize the pattern of these scans and prevent industrial control systems from being indexed by the powerful Shodan tool.

Hadi Fazelinia
(M.Sc.)

Adversarial Attacks against Dynamic Analysis based Malware Detection Systems

Malware detection systems are one of the important needs of every organization. In this research, malware detection methods, used techniques, shortcomings and strengths of these tools are investigated and finally, a model is designed so that we can evade Dynamic Analysis based Malware Detection Systems. Our ultimate goal is to improve malware detection systems.

Homa Shirafkan
(M.Sc.)

Dynamic selection of cloud data center optimization objective

The purpose of data center optimization can be different (e.g., energy consumption, resource efficiency, energy cost, SLA). In most data centers, these goals are fixed. In this research, we intend to have these goals dynamically and based on a set of KPI in a data center.
1- Adaptive virtual machine placement: a dynamic approach for energy-efficiency, QoS enhancement, and security optimization

Ebrahim Farjamfard
(M.Sc.)

Anomaly Detection Using Conditional Generative Adversarial Networks and Transformer Based on System Logs

Diagnosing the anomaly of log-based systems with high accuracy is one of the important needs of the industry so that they can quickly return their services to their original state. In this research, we intend to use the transformer model, which has a good performance in a long sequence of texts, regarding logs. We also use the CGAN model to eliminate the imbalance between different classes of anomalies.

Amir Mohammad Karamzadeh
(M.Sc.)

Cold Start Latency Mitigation in Serverless Computing

One of the services in cloud computing is the computational model of serverless computing. In serverless computing, user codes are hosted without knowing the details of the infrastructure. One of the challenges with this model is the cold start problem. In this project, we intend to reduce the cold start delay.
1- Reducing cold start delay in serverless computing using lightweight virtual machines

Faezeh Montazerin
(M.Sc.)

A fluctuations-in-resource-demand aware virtual machine placement algorithm for cloud data centers

Reducing energy consumption in data centers is one of the major challenges. In most of the past work, fluctuations in user requests in resources in the location of virtual machines are not considered. In this research, we intend to optimize the location of the virtual machine, taking into account changes in resource demand, in order to reduce migration and optimize the energy consumption of the centers.
1- Enhancing virtual machine placement efficiency in cloud data centers through fluctuations-aware resource management

Arash Hadadi
(M.Sc.)

Intelligent countermeasure selection for Information security risk assessment

Virtual machine placement in the cloud environment can be done with different approaches. In this research, the maintenance of service quality for a distributed application in the cloud environment is considered.
1-A Quality of Service Aware VM Placement for User Applications in Cloud Data Center

Negar Baharvand
(M.Sc.)

Service Function Chain Scheduling Problem in Edge Computing

Service function chain scheduling, which is based on virtualization of network functions, has a significant impact on system performance and service delivery to users in the cloud. In this research, we decided to address the problem of service function chain scheduling in edge networks by considering two improvements: 1) network reconfiguration and 2) optimization of network resources.
1- Optimizing service function chaining and scheduling with deep Q-learning for enhanced quality of experience in edge networks

Amin Bagheri
(M.Sc.)

Automating the translation of high-level security requirements for cloud users into an optimal placement model

Optimal placement of security functions in the cloud by cloud providers is one of the major challenges. The reason is to translate the user needs to the placement model. In this project, we intend to perform this challenge automatically.
1- Automating the Translation of Cloud Users' High-Level Security Needs to an Optimal Placement Model in the Cloud Infrastructure

Mohammad Osmanpoor
(M.Sc.)

Energy efficient workflow scheduling in cloud computing environment

The growth of energy consumption has been explosive in current data centers. This explosion has led to greater advocacy of green computing, and many efforts and works focus on the task scheduling in order to reduce energy dissipation. In order to obtain more energy reduction as well as maintain the quality of service, this research proposes a DVFS-enabled Energy-efficient Workflow Scheduling algorithm.
1- Convergence of the Harris hawks optimization algorithm and fuzzy system for cloud-based task scheduling enhancement

Parisa Kalaki
(M.Sc.)

Anomaly Detection on OpenStack Log Analysis

OpenStack is a free open-source cloud computing platform that includes common services such as computing, image, network, and storage services. OpenStack Log Analysis is a significant part of detecting anomaly behavior. In this research, we analyze OpenStack log and divide it into two sections, log parsing and log mining. Furthermore, different datamining algorithms are used to find any unusual patterns that cause faults.
1- Anomaly detection on OpenStack logs based on an improved robust principal component analysis model and its projection onto column space

Ehsan Rasoulpour
(M.Sc.)

 intelligent VM placement method for minimizing energy cost and carbon emission in distributed cloud data centers

By considering energy and carbon efficient virtual machine (VM) placement approaches, not only would the total power usage and carbon emission be reduced, but the excessive amounts of data centers' expenses including energy cost and carbon tax would also be significantly declined. In this research, for overcoming this compelling issue, we introduce a novel, intelligent VM placement approach for reducing energy consumption, carbon emission, and their related costs.

Mohammad Khanahmadi
(M.Sc.)

Performance Analysis in Microservice Architecture using OpenTracing

A microservice based cloud program consists of several services, each developed separately. OpenTracing is a new standard for controlling and viewing in-app events at different levels in the cloud. In this research, we intend to use OpenTracing by calculating multi metrics such as throughput, response time, request ratio of success and error to analyze the performance of software implemented on a cloud scale and will use the results obtained to identify system bottlenecks and monitor the system trends closely.
1- Detection of microservice‐based software anomalies based on OpenTracing in cloud

Sina Tamjidy
(M.Sc.)

Intelligent countermeasure selection for Information security risk assessment

Accordion SaThese days, the risk assessment of an organization's assets is a challenging task. If organizations did not pay significant attention to the security of their data and an attack happens against them, they will face too many problems like reputation loss or legal problems. One of the best ways to improve the process of risk assessment is by using our knowledge about data extracted from the business processes. With this kind of framework, we can make an accurate risk assessment model. What is left to do is using the logs of this system to selecting the mitigations based on the previous system's state in an intelligent way, which is the subject of this research.mple Description
1- Intelligence in security countermeasures selection

Alireza Khajouei
(M.Sc.)

Incremental VM placement

Over the last few years, virtual machine placement has become a chief operation in cloud computing. Researchers confront a lot of challenges according to different aspects such as computational complexity, end-to-end delay and etc.. in our placement, we want to model the optimal solution by integer linear programming and represent a greedy algorithm to make an incremental placement independent of the cost function.
1- A Scalable Stateful Approach for Virtual Security Functions Orchestration

Sara Azizi
(M.Sc.)

Improving cvss vulnerability scoring

Ac CVSS is one of the most popular methods for calculating vulnerability severity. One of the disadvantages of the CVSS method is the failure to use the vulnerability history, which is addressed in this research. In the history of vulnerability, a variety of metrics can be considered, for example, when the vulnerability is fixed, the number of attacks between detection and fixing. The purpose of this research is to focus on these metrics.cordion Sample Description

Puya Pakshad
(M.Sc.)

A security test oracle based on security features extracted from source-code

The importance of software application's security has increased rapidly over the recent years. Software security testing helps us to check the security behavior of a program. To determine whether the program behaves securely, we need an oracle who detects the correct program's behavior from security concept. There are various approaches to construct security test oracles. In this research, we will present a security test oracle based on security features which are extracted from source code of program to identify vulnerable codes and their types.
1- A security vulnerability predictor based on source code metrics

Niloofar Moradi
(M.Sc.)

Incremental NVF Optimal Placement

Network Function Virtualization (NFV) is a paradigm to facilitate dynamic provisioning of network services through virtualization technologies. The optimal placement of virtual network functions is known to be an NP-hard problem. Our placement objectives consist of three goals: 1) minimizing end-to-end delay for each flow, 2) maximizing the network resource utilization, 3) minimizing network function installation costs.
1- A Scalable Stateful Approach for Virtual Security Functions Orchestration

Sima Bagheri
(M.Sc.)

Firewall: Scaling up and down in cloud

Firewalls have been widely deployed for securing networks and are the first line for defending against malicious traffic. A firewall checks each incoming or outgoing packet to decide whether to accept or reject the packet based on its policy. Setting up a centralized firewall for a whole cloud data center is infeasible. In this project, we aim at addressing most of the actual challenges in the centralized firewall by outsourcing some parts of the main firewall as several small firewalls.
1- Dynamic Firewall Decomposition and Composition in the Cloud.

Ehsan Khodayarseresht
(M.Sc.)

Reducing Energy and Carbon Cost of Multiple Cloud Providers Using Cloud Federation and Solar Energy Prediction

Nowadays, the highest amount of money paid by cloud providers is about energy and carbon cost. The computation of energy and carbon cost plays a crucial role in data centers' operational costs and also could help to reduce their carbon footprint rate. In this research, generally, we aim to reduce the overall energy and carbon cost of multiple cloud providers with the use of cloud federation, renewable energy and other effective issues such as idle host awareness, solar energy prediction and dynamic PUE.
1- Reducing Energy and Carbon Cost of Cloud Data Centers through Optimal Virtual Machine Placement algorithms and an Automatic Servers Scale Controller Model

2- A multi-objective cloud energy optimizer algorithm for federated environments

3- Energy and carbon-aware initial VM placement in geographically distributed cloud data centers
Former Students