Baris Yamansavascilar

Failures in networks result in service disruptions which may cause deteriorated Quality of Service (QoS) for the end-users. Since SDN is becoming the mainstream paradigm for networks, the implementation of a robust fault tolerance scheme for SDN-based networks is crucial. Moreover, that robust fault tolerance scheme should ensure the stability for QoS and Quality of Experience (QoE). To this end, we investigated

• A novel dynamic protection method considering the quality of alternative paths.

• How video quality and thus QoE would fluctuate when a failure occurs.

• The effect of the detection of the congestion in the link layer operating Bidirectional Forwarding Detection (BFD) rather than the application layer.

• The impact of BFD intervals on the QoE parameters along with the video segment size to come up with a feasible operational range where the fault tolerance approach is beneficial for the congestion case.

Note that this study was my thesis subject in my MS studies. 

The improvements in the edge computing technology pave the road for diversified applications that demand real-time interaction. However, due to the mobility of the end-users and the dynamic edge environment, it becomes challenging to handle the task offloading with high performance. Moreover, since each application in mobile devices has different characteristics, a task orchestrator must be adaptive and have the ability to learn the dynamics of the environment. For this purpose, we develop a deep reinforcement learning based task orchestrator, DeepEdge, which learns to meet different task requirements without needing human interaction even under the heavily-loaded stochastic network conditions in terms of mobile users and applications. Given the dynamic offloading requests and time-varying communication conditions, we successfully model the problem as a Markov process and then apply the Double Deep Q-Network (DDQN) algorithm to implement DeepEdge. To evaluate the robustness of DeepEdge, we experiment with four different applications including image rendering, infotainment, pervasive health, and augmented reality in the network under various loads. Furthermore, we compare the performance of our agent with the four different task offloading approaches in the literature. Our results show that DeepEdge outperforms its competitors in terms of the percentage of satisfactorily completed tasks.

The project aimed at detecting and preventing Intrusion and Advanced Persistent Thread (APT) attacks via manifesting anomalies in network flows and system logs using machine learning algorithms. DeterLab Testbed was used as a development and test platform for the various work packages of the project. Moreover, it was funded by The Scientific and Technological Research Council of Turkey (TUBITAK).

Recent developments on the Internet have led to increase the importance of network traffic classification. Classification of large-scale network traffic can provide much important information for network management, QoS and network security. In this study, the influence of different network speeds to accuracy was evaluated and a comparison between packet-based and time-based approaches was performed.

Great hardware and software capabilities of mobile devices allow us to research new scientific fields. Activity recognition is one of the main research areas for smartphones. Built-in sensors of a standard smartphone, such as accelerometer, magnetometer, gyroscope, enable us to predict the daily activities of a person. However, continuous samplings from each sensor increase the CPU utilization and cause the battery to die sooner. In this study, we focused on the window sizes and the sampling rates in order to observe how they affect to the accuracy and CPU utilization.

Failures in networks result in service disruptions which may cause deteriorated Quality of Service (QoS) for the end-users. Since SDN is becoming the mainstream paradigm for networks, the implementation of a robust fault tolerance scheme for SDN-based networks is crucial. Moreover, that robust fault tolerance scheme should ensure the stability for QoS and Quality of Experience (QoE). To this end, we investigated

• A novel dynamic protection method considering the quality of alternative paths.

• How video quality and thus QoE would fluctuate when a failure occurs.

• The effect of the detection of the congestion in the link layer operating Bidirectional Forwarding Detection (BFD) rather than the application layer.

• The impact of BFD intervals on the QoE parameters along with the video segment size to come up with a feasible operational range where the fault tolerance approach is beneficial for the congestion case.

Note that this study was my thesis subject in my MS studies. 

The improvements in the edge computing technology pave the road for diversified applications that demand real-time interaction. However, due to the mobility of the end-users and the dynamic edge environment, it becomes challenging to handle the task offloading with high performance. Moreover, since each application in mobile devices has different characteristics, a task orchestrator must be adaptive and have the ability to learn the dynamics of the environment. For this purpose, we develop a deep reinforcement learning based task orchestrator, DeepEdge, which learns to meet different task requirements without needing human interaction even under the heavily-loaded stochastic network conditions in terms of mobile users and applications. Given the dynamic offloading requests and time-varying communication conditions, we successfully model the problem as a Markov process and then apply the Double Deep Q-Network (DDQN) algorithm to implement DeepEdge. To evaluate the robustness of DeepEdge, we experiment with four different applications including image rendering, infotainment, pervasive health, and augmented reality in the network under various loads. Furthermore, we compare the performance of our agent with the four different task offloading approaches in the literature. Our results show that DeepEdge outperforms its competitors in terms of the percentage of satisfactorily completed tasks.

This project aimed at implementing an MU MAC OFDMA scheduler on an AP that supports IEEE 802.11ax standard (Wifi 6). The expectation from the scheduler was to consider the fairness and efficiency between HE STAs regarding their uplink and downlink traffic. The project was funded by AirTies Wireless Networks.

Failures in networks result in service disruptions which may cause deteriorated Quality of Service (QoS) for the end-users. Since SDN is becoming the mainstream paradigm for networks, the implementation of a robust fault tolerance scheme for SDN-based networks is crucial. Moreover, that robust fault tolerance scheme should ensure the stability for QoS and Quality of Experience (QoE). To this end, we investigated

• A novel dynamic protection method considering the quality of alternative paths.

• How video quality and thus QoE would fluctuate when a failure occurs.

• The effect of the detection of the congestion in the link layer operating Bidirectional Forwarding Detection (BFD) rather than the application layer.

• The impact of BFD intervals on the QoE parameters along with the video segment size to come up with a feasible operational range where the fault tolerance approach is beneficial for the congestion case.

Note that this study was my thesis subject in my MS studies. 

Recent developments on the Internet have led to increase the importance of network traffic classification. Classification of large-scale network traffic can provide much important information for network management, QoS and network security. In this study, the influence of different network speeds to accuracy was evaluated and a comparison between packet-based and time-based approaches was performed.

This project aimed at implementing an MU MAC OFDMA scheduler on an AP that supports IEEE 802.11ax standard (Wifi 6). The expectation from the scheduler was to consider the fairness and efficiency between HE STAs regarding their uplink and downlink traffic. The project was funded by AirTies Wireless Networks.