The detection of anomalies in Computer Networks is one of the most considerable challenges that experts in this field are facing nowadays. Thus far, different artificial intelligence methods and algorithms have been proposed, tested, and utilized for detecting these anomalies. However, attempts made to enhance the speed and accuracy of these anomalies’ detection process are continuously ongoing. In this research, pattern recognition based on artificial neural Networks is applied to automatically detect anomalies in Computer Networks. Also, to increase the speed of the pattern recognition based on the process of the neural network, the principal component analysis algorithm will be used as a method for dimension reduction of training samples. The results of the performed simulations based on the proposed methods in this research show that dimension reduction of training samples by principal component analysis algorithm and then applying the pattern recognition based on neural Networks method leads to high-speed (less than 10 seconds) and high-accuracy (99-100%) detection of anomalies in Computer Networks.

Today, the number of alerts issued by network security systems has increased significantly and network administrators encounter new problems in handling the issued alerts and responding to them. As managing and responding to such a large number of alerts is difficult, alert management and intrusion response system (IRS) are the main part of the security protection systems including intrusion detection systems. The main task of alert management is to reveal the attack details to IRS. Subsequently, the appropriate responses are applied to reduce the attack damage and recover the compromised Computer Networks back to their normal operational mode. In the literature, researchers have investigated alert management techniques and IRS solutions separately, despite the fact that alert management is one of the basic requirements of response process and its outcome directly affects the IRS performance. Alert management design should provide the necessary information about the attacks to the response system according to its type and requirements. This information along with information from network resources present the current state of the network to IRS. However, if decisions taken by the response system is only based on the current network status, the total cost of the network will increase over the time. Therefore with a futuristic concept and considering the present available information and all possible coming states, decision making process in the response system can be improved. In this paper, using a futuristic approach we seek to propose optimal solutions for confronting already-occurred and future-probable attacks. To achieve this goal, the proposed framework contains two subsystems: attacks and alerts modeling, and response modeling. In the first subsystem, we analyze the IDS alerts to find the similarity and causality relationships. We also present a comprehensive approach for network attack forecasting to obtain some useful predictions about the future states of the network. In the second subsystem, the response analyzer presents a multilevel response model to categorize intrusion responses. It also provides a foresight model to estimate the response cost by considering IDS alerts, network dependencies, attack damage, response impact, and the probability of potential attacks. Finally, models are proposed to make the best decision based on available information about the present and all possible coming states. Simulation results for different scenarios show that the response system, with a prospective vision, steers the network toward desired states with reduced cost of attack and response.

In security risk management of Computer Networks, some challenges are more serious in large Networks. Specifying and estimating risks is largely dependent on the knowledge of security experts. In this paper, a framework for security risk estimation is proposed to address this issue. It represents the security knowledge required for security risk estimation and utilizes current security metrics and vulnerability databases. This framework is a major step towards automating the process of security risk estimation so that a network administrator can estimate the risk of the network with less expertise and e ort. As a case study, the proposed framework is applied to a sample network to show its applicability and usability in operational environments. The comparison of results with two existing methods showed the validity of the estimations given by the proposed framework.

In junctionless nanowire transistors, drain-channel-source doping is of the same type and level. Therefore, fabricating junctionless transistors is less complicated than inversion mode transistors. However, the reduced ratio of on current to off-state current (ION/IOFF) in junctionless nanowire transistors has made their operation difficult due to the high impurity scattering in the channel. The larger ION / IOFF ratio indicates an increase in transistor switching speed. In this study, the use of gate oxide engineering is proposed to increase of ION / IOFF ratio. The proposed oxide thickness is 1.5nm, which is a buffer layer with K = 5.7 and a thickness of 0.5nm, and the oxide with a high permittivity factor with K = 29 and a thickness of 1nm. The proposed structure is called GAA-JL-FET-with-oxide buffer layer. In GAA-JL-FET-with-oxide buffer layer, the ION / IOFF ratio has been improved by about 106 and 102 times compared to the structure with K = 3.9 and K = 7.5, respectively. Due to the high switching speed, the GAA-JL-FET-with-oxide buffer layer is applied to design the multiple-valued logic (MVL) interconnection network in the voltage mode for the first time using the mixed-mode tool in Silvaco software. The results reveal improving subthreshold parameters of the simulated device significantly enhance the average power, max delay and power delay product (PDP) of the proposed interconnection Networks.