Round Trip Time Research Articles

Performance and improvements of TCP CUBIC in low-delay cellular networks

Growing Internet usage on mobile devices in cellular networks makes it increasingly important to assure good performance for web traffic, which often consists of many short-lived TCP flows. For TCP, and in the future likely also for QUIC, CUBIC is the most widely deployed congestion control algorithm. With Linux, HyStart is used as default Slow Start algorithm. In this paper we investigate performance impairments occurring when the combination of CUBIC and HyStart is used in low-delay cellular networks. To explain the problem in detail, we present measurements of characteristic per-packet round-trip times caused by uplink scheduling in mobile networks. Inherent round-trip time variations often trigger an early exit of Slow Start, causing CUBIC to underutilize the available capacity. In addition, we extend earlier throughput models for TCP CUBIC to accurately estimate the throughput for short-lived flows as a function of the congestion window size at the Slow Start exit point. We validate the model by measurement results. Further, we propose two HyStart improvements enhancing TCP CUBIC’s performance significantly. In our LTE testbed the mean throughput of a 5MB data transfer increases by 10.5%. Given the nature of web traffic, these fixes can improve the web experience of users noticeably. This shows the importance of measurements, testing, and subsequent optimizations when designing and evaluating transport and congestion control mechanisms for mobile networks.

An OCF-IoTivity enabled smart-home optimal indoor environment control system for energy and comfort optimization

The growing energy demand with diminishing energy resources calls for development of optimal indoor control system through accurate energy and occupant comfort modeling in a smart-home. While some major obstacles faced by smart-home IoT network are heterogeneity and disparate IP frameworks resulting in a connectivity and interoperability issues. In this paper we developed a new Open Connectivity Foundation (OCF) based prediction assisted optimal control framework for optimizing energy consumption and maximizing occupant comfort in smart-home. To the best of our knowledge, we are the first to design a scalable, secure, and inter-operable optimal control solution for smart-home IoT networks. The developed framework offloads machine learning models on IoT device for accurate energy and thermal comfort modeling to enable prediction assisted optimization. The system enables edge analytics using deep learning based inference models for proactive response. For optimization standard Firefly algorithm is modified based on inertia weight approach to achieve improved performance. OCF based optimal actuator control test-bed deployment and real-time experimentation is conducted to evaluate the performance of the system. Empirical investigation verified the effectiveness of proposed framework with energy saving scope of 36.82%, 5 s average response time and 3.36 MS Round Trip Time.

A complete architecture for Ambient Assisted Living scenarios using a cross protocol proxy

Nowadays, in the most developed countries, modern society is moving towards scenarios in which the self-sufficiency elderly people live alone in their homes. An automatic remote monitoring system using wearable sensors is becoming even more important in Ambient Assisted Living applications. For this type of services, it is important that IoT sensors networks, which are generally composed of devices with limited computing power and storage, implement reliable communication among sensors and the Internet. There are several specialized protocols for the Internet of Things proposed by the scientific community, each characterized by its own levels of Quality of Services. The emergence of new protocols forces the need for developing proxying systems able to intermediate among different types of networks and to translate the relative protocols. In this paper, we propose a complete architecture for monitoring and managing wearable devices, and, in particular, fall detection ones. Our system uses a cross protocol proxy and a device with CoAP and MQTT as application level protocols, while it exploits the NB-IoT at physical and data-link levels. The goal of this work is the performance evaluation of the proposed solution in terms of Throughput, Round Trip Time and Delay. The results highlight the low latency reached by the proposed system architecture thanks to the implemented protocols.

CloudPilot: Flow acceleration in the cloud

TCP-split proxies have been previously studied as an efficient mechanism to improve the rate of connections with large round trip times. These works focused on improving a single flow. In this paper, we investigate how strategically deploying TCP-split proxies in the cloud can improve the performance of geo-distributed applications entailing multiple flows interconnecting globally-distributed sources and destinations using different communication patterns, and being subject to budget limitations.We present CloudPilot, a Kubernetes-based system that measures communication parameters across different cloud regions, and uses these measurements to deploy cloud proxies in optimized locations on multiple cloud providers. To this end, we model cloud proxy acceleration and define a novel cloud-proxy placement problem. Since this problem is NP-Hard, we suggest a few efficient heuristics to solve it. Finally, we find that our cloud-proxy optimization can improve flow completion time by an average of 3.6× in four different use cases.

PBQ-Enhanced QUIC: QUIC with Deep Reinforcement Learning Congestion Control Mechanism.

Currently, the most widely used protocol for the transportation layer of computer networks for reliable transportation is the Transmission Control Protocol (TCP). However, TCP has some problems such as high handshake delay, head-of-line (HOL) blocking, and so on. To solve these problems, Google proposed the Quick User Datagram Protocol Internet Connection (QUIC) protocol, which supports 0-1 round-trip time (RTT) handshake, a congestion control algorithm configuration in user mode. So far, the QUIC protocol has been integrated with traditional congestion control algorithms, which are not efficient in numerous scenarios. To solve this problem, we propose an efficient congestion control mechanism on the basis of deep reinforcement learning (DRL), i.e., proximal bandwidth-delay quick optimization (PBQ) for QUIC, which combines traditional bottleneck bandwidth and round-trip propagation time (BBR) with proximal policy optimization (PPO). In PBQ, the PPO agent outputs the congestion window (CWnd) and improves itself according to network state, and the BBR specifies the pacing rate of the client. Then, we apply the presented PBQ to QUIC and form a new version of QUIC, i.e., PBQ-enhanced QUIC. The experimental results show that the proposed PBQ-enhanced QUIC achieves much better performance in both throughput and RTT than existing popular versions of QUIC, such as QUIC with Cubic and QUIC with BBR.

Time indices of pre-hospital EMS missions before and during the COVID-19 pandemic: a cross-sectional study in Iran

BackgroundThe COVID-19 pandemic resulted in many changes in pre-hospital emergency medical services (EMS), including wearing full-body protective suits and well-fitted face masks, which can influence time indices in the course of service delivery. The present study aimed to compare the time indices of pre-hospital EMS missions before and during the COVID-19 pandemic in Iran.MethodsThis descriptive cross-sectional study used census sampling to select 17,860 emergency calls that caused patient transfer to medical facilities from March 2018 to March 2021 and then examined the time indices of pre-hospital EMS missions during the COVID-19 pandemic. The data collection tool was a two-part checklist: patients' individual characteristics and pre-hospital EMS mission time indices. The data were further analyzed using the SPSS16 and independent samples t-test.ResultsOut of the patients transferred by the EMS, 11,773 cases (65.9%) were male and 6,087 (34.1%) were female. The most common reason for the emergency calls was accidents (28.0%). Moreover, response time (P < 0.001), on-scene time (P < 0.001), transfer time (P < 0.001), total run time (P < 0.001), and round trip time (P < 0.001) increased significantly during the COVID-19 pandemic.ConclusionsWe concluded that the EMS time indices elevated following the COVID-19 pandemic. Updating pre-hospital information management systems, ambulances and medical equipment, as well as holding training courses for pre-hospital EMS personnel could effectively enhance the time indices of pre-hospital EMS missions.

Using Metaheuristics (SA-MCSDN) Optimized for Multi-Controller Placement in Software-Defined Networking

The multi-controller placement problem (MCPP) represents one of the most challenging issues in software-defined networks (SDNs). High-efficiency and scalable optimized solutions can be achieved for a given position in such networks, thereby enhancing various aspects of programmability, configuration, and construction. In this paper, we propose a model called simulated annealing for multi-controllers in SDN (SA-MCSDN) to solve the problem of placing multiple controllers in appropriate locations by considering estimated distances and distribution times among the controllers, as well as between controllers and switches (C2S). We simulated the proposed mathematical model using Network Simulator NS3 in the Linux Ubuntu environment to extract the performance results. We then compared the results of this single-solution algorithm with those obtained by our previously proposed multi-solution harmony search particle swarm optimization (HS-PSO) algorithm. The results reveal interesting aspects of each type of solution. We found that the proposed model works better than previously proposed models, according to some of the metrics upon which the network relies to achieve optimal performance. The metrics considered in this work are propagation delay, round-trip time (RTT), matrix of time session (TS), average delay, reliability, throughput, cost, and fitness value. The simulation results presented herein reveal that the proposed model achieves high reliability and satisfactory throughput with a short access time standard, addressing the issues of scalability and flexibility and achieving high performance to support network efficiency.

Towards secure asynchronous messaging with forward secrecy and mutual authentication

Instant messaging applications like WhatsApp and iMessage have been widely used to exchange public and private information for both individuals and organizations. At the same time, the asynchronicity nature of these applications also introduces various security issues. To this end, considerable efforts have been made to strengthen the security and privacy of these applications, where a major goal is to capture the forward secrecy of asynchronous messaging without sacrificing other reasonable security properties. In this paper, we mainly focus on forward-secure zero round-trip time (0-RTT) key exchange protocols. Such a protocol enables a user to send the encrypted application data along with the first protocol message, and thus is especially suitable for securing asynchronous messaging. Concretely, we introduce a new cryptographic primitive dubbed as puncturable identity-based matchmaking key encapsulation mechanism (PIB-MKEM). It captures the authenticity of ciphertexts, and also allows a receiver to repeatedly update his/her secret key with a puncture algorithm. Based on PIB-MKEM, we further put forth a generic construction of 0-RTT key exchange protocol that enjoys forward secrecy and mutual authentication as well as replay resistance in parallel. We also present a concrete construction of PIB-MKEM, which immediately instantiates the above generic 0-RTT key exchange protocol. Moreover, we implement the proposed PIB-MKEM construction, and provide extensive evaluations to demonstrate its practicability.

A Novel Deep-Learning Model for Remote Driver Monitoring in SDN-Based Internet of Autonomous Vehicles Using 5G Technologies

The rapid advancement in the Internet of Things (IoT) and its integration with Artificial Intelligence (AI) techniques are expected to play a crucial role in future Intelligent Transportation Systems (ITS). Additionally, the continuous progress in the industry of autonomous vehicles will accelerate and increase their short adoption in smart cities to allow safe, sustainable and accessible trips for passengers in different public and private means of transportation. In this article, we investigate the adoption of 5G different technologies, mainly, the Software-Defined Networks (SDN) to support the communication requirements of delegation of control of level-2 autonomous vehicles to the Remote-Control Center (RCC) in terms of ultra-low delay and reliability. This delegation occurs upon the detection of a drowsy driver using our proposed deep-learning-based technique deployed at the edge to reduce the level of accidents and road congestion. The deep learning-based model was evaluated and produced higher accuracy, precision and recall when compared to other methods. The role of SDN is to implement network slicing to achieve the Quality of Service (QoS) level required in this emergency case. Decreasing the end-to-end delay required to provide feedback control signals back to the autonomous vehicle is the aim of deploying QoS support available in an SDN-based network. Feedback control signals are sent to remotely activate the stopping system or to switch the vehicle to direct teleoperation mode. The mininet-WiFi emulator is deployed to evaluate the performance of the proposed adaptive SDN framework, which is tailored to emulate radio access networks. Our simulation experiments conducted on realistic vehicular scenarios revealed significant improvement in terms of throughput and average Round-Trip Time (RTT).

Demonstration of deep-space synchronous two-way laser ranging with a laser transponder aboard Hayabusa2

In December 2020, synchronous two-way laser ranging in deep space was successful for the first time over distances up to 6.46 million km. A laser altimeter aboard spacecraft Hayabusa2 was used as an optical transponder, detecting laser pulses emitted from the ground stations on Earth and retransmitting the pulses from the spacecraft to the ground. These retransmitted pulses were successfully detected at the ground station. The experiment was conducted as a demonstration of the deep-space laser ranging. The repetition frequency of the onboard instrument was limited to 0.5 Hz, and there was a lot of background noise because the experiments were carried out in daytime on the ground. Nevertheless, laser detection on the ground was still possible due to the high time-correlation between the detected and the predicted in the presence of random noise. Procedures similar to satellite laser ranging were applied to create residuals by subtracting the orbit predictions of the spacecraft from the observed round-trip time. Furthermore, some of the transponder return pulses were uniquely identified, where the time coherence of the measurements had been recovered from the telemetry data of the laser altimeter.

1.3-μm 2.8-ns electro-optical cavity damping Nd:YVO&lt;sub&gt;4&lt;/sub&gt; laser

&lt;sec&gt;1.3-μm Nd laser has significant practical applications in various fields, such as fiber communication, medical treatment, frequency conversion, and scientific research. Many applications of a 1.3-μm laser, particularly frequency conversion, benefit greatly from a short pulse width with high peak power. In the paper, an electro-optical cavity dumping Nd:YVO&lt;sub&gt;4&lt;/sub&gt; laser at 1342 nm wavelength is studied theoretically and experimentally.&lt;/sec&gt;&lt;sec&gt;The pulse width for an electro-optical cavity dumping laser is determined by the optical length of the cavity. A narrower pulse width is obtained by reducing the length of the cavity and the round trip time of the laser in the cavity. However, when the round trip time in the cavity approaches to the falling edge time of the electro-optical switch, shortening the length of the cavity will not obtain a narrower pulse width, and the falling edge time of the electro-optical switch will influence the laser pulse width. The temporal characteristics of the laser pulse are simulated when the falling edge time of the electro-optical switch is close to the round trip time in the cavity.&lt;/sec&gt;&lt;sec&gt;The influence of the falling edge time of the electro-optical switch on the laser pulse duration is analyzed theoretically. The modified rate equation is used to study the relationship between the falling edge time and the laser pulse width.&lt;/sec&gt;&lt;sec&gt;We demonstrate an electro-optical cavity dumping Nd:YVO&lt;sub&gt;4&lt;/sub&gt; laser. The atom percent of 0.3% Nd:YVO&lt;sub&gt;4&lt;/sub&gt; placed in a short Plano-concave cavity is in-band pumped by an 880 nm quasi-continuous-wave diode. A fiber-coupled diode laser module (&lt;i&gt;NA&lt;/i&gt; = 0.22) with a power of 30 W is used. An LiNbO&lt;sub&gt;3&lt;/sub&gt; electro-optical switch is employed for the cavity-dumping. The 1342-nm cavity-dumping laser operates at a repetition rate of 1 kHz, single-pulse energy of 0.21 mJ, and pulse width of 2.8 ns. Near-diffraction-limited beam quality with an &lt;inline-formula&gt;&lt;tex-math id="Z-20230713064624-1"&gt;\begin{document}$ M^2 $\end{document}&lt;/tex-math&gt;&lt;alternatives&gt;&lt;graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="14-20230014_Z-20230713064624-1.jpg"/&gt;&lt;graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="14-20230014_Z-20230713064624-1.png"/&gt;&lt;/alternatives&gt;&lt;/inline-formula&gt;value of &lt; l.2 is achieved. The setup uses MgO:PPLN crystal to generate efficient second harmonic at 671 nm, with a pulse width of 1.8 ns. To the best of our knowledge, this is the shortest pulse duration ever obtained from 1.3 μm actively &lt;i&gt;Q&lt;/i&gt;-switched Nd-doped laser.&lt;/sec&gt;

1342 nm 2.8ns electro-optical cavity damping Nd: YVO&lt;sub&gt;4&lt;/sub&gt; laser

1.3 μm Nd laser has significant practical applications in various fields, such as fiber communication, medical treatment, frequency conversion, and scientific research, etc. Many applications of a 1.3 μm laser, especially like frequency conversion, benefit much from a short pulse width with high peak power. In the paper, an electro-optical cavity dumping Nd:YVO&lt;sub&gt;4&lt;/sub&gt; laser at 1342 nm wavelength has been studied theoretically and experimentally. The pulse width for an electro-optical cavity dumping laser is determined by the optical length of the cavity. A narrower pulse width can be obtained by reducing the length of the cavity and the round trip time of the laser in the cavity. However, when the round trip time in the cavity approaches to the falling edge time of the electro-optical switch, shortening the length of the cavity will not get a narrower pulse width, and the falling edge time of the electro-optical switch will influence the laser pulse width. The temporal characteristics of the laser pulse are simulated when the falling edge time of the electro-optical switch is close to the round trip time in the cavity. Influences of the falling edge time of the electro-optical switch on the laser pulse duration are analyzed theoretically. The modified rate equation is used to study the relationship between the falling edge time and the laser pulse width. We demonstrate an electro-optical cavity dumping Nd:YVO&lt;sub&gt;4&lt;/sub&gt; laser. The 0.3 at.% Nd:YVO&lt;sub&gt;4&lt;/sub&gt; placed in a short Plano-concave cavity is in-band pumped by an 880 nm quasi-continuous-wave diode. A fiber-coupled diode laser module (NA=0.22) with a power of 30 W was used. A LiNbO&lt;sub&gt;3&lt;/sub&gt; electro-optical switch was employed for the cavity-dumping. The 1342 nm cavity-dumping laser operates at a repetition rate of 1kHz, and a single-pulse energy of 0.21 mJ is obtained with a pulse width of 2.8 ns. Near-diffraction-limited beam quality with an M&lt;sup&gt;2&lt;/sup&gt; value of &lt; l.2 is achieved. The setup provides efficient second harmonic generation at 671 nm using a MgO:PPLN crystal, and the pulse width is 1.8 ns. To the best of our knowledge, this is the shortest pulse duration obtained from 1.3μm actively Q-switched Nd-doped laser.

Reliable Failure Restoration with Bayesian Congestion Aware for Software Defined Networks

Software Defined Networks (SDN) introduced better network management by decoupling control and data plane. However, communication reliability is the desired property in computer networks. The frequency of communication link failure degrades network performance, and service disruptions are likely to occur. Emerging network applications, such as delay-sensitive applications, suffer packet loss with higher Round Trip Time (RTT). Several failure recovery schemes have been proposed to address link failure recovery issues in SDN. However, these schemes have various weaknesses, which may not always guarantee service availability. Communication paths differ in their roles; some paths are critical because of the higher frequency usage. Other paths frequently share links between primary and backup. Rerouting the affected flows after failure occurrences without investigating the path roles can lead to post-recovery congestion with packet loss and system throughput. Therefore, there is a lack of studies to incorporate path criticality and residual path capacity to reroute the affected flows in case of link failure. This paper proposed Reliable Failure Restoration with Congestion Aware for SDN to select the reliable backup path that decreases packet loss and RTT, increasing network throughput while minimizing post-recovery congestion. The affected flows are redirected through a path with minimal risk of failure, while Bayesian probability is used to predict post-recovery congestion. Both the former and latter path with a minimal score is chosen. The simulation results improved throughput by (45%), reduced packet losses (87%), and lowered RTT (89%) compared to benchmarking works.

TCP-Puerto-Londero: A New Approach for End-to-End Queue Length Control

This paper presents a new TCP protocol called TCPPuerto-Londero, which makes dynamic tuning in the congestion window (cwnd) by means of adaptive control theory aiming to keep stable and small the queue length in the bottleneck link located on the path from source to destination. This adaptive control loop has the relative delay in the forward path measured as an input and the cwnd value as an output. Thus, unlike classic TCP protocols, TCP-Puerto-Londero does not use Round Trip Time (RTT) information. Also, unlike classic Active Queue Management (AQM) strategies and Explicit Congestion Notification (ECN) based protocols, TCP-Puerto-Londero employs end-to-end queue management without the need for ECN resources. Moreover, TCP-Puerto-Londero also aims to attend to the new challenges of the Industrial Internet of Things (IIoT). Its performance has been tested in a Dumbbell network topology shared by TCP and UDP-like Networked Control Systems (NCS) flows. Therefore, TCP-Puerto-Londero performance has been compared with ECN-based protocols like DCTCP, E-DCTCP, TCP-Jersey, and ENCN. Furthermore, this paper employs an approach for modeling, analysis, simulation, and verification of the communication network and NCS employing UPPAAL simulation software tool, where all network constituents (transmitters, channels, routers, receivers, controllers, and plants) were modeled employing timed automata, simplifying a formal verification of the complete studied system. Simulations and statistical verification indicate that even though utilizing fewer resources (as it does not require the AQM/ECN information) TCP-Puerto-Londero overcomes TCP-Jersey, DCTCP, and EDCTCP with regard to throughput and fairness for TCP flows. Also, TCP-Puerto-Londero flows are capable to keep the queue length more stable and smaller than other protocols that were compared, and consequently, it reduces the impact in NCS-UDPlike flows sharing the same network, whose performance was measured employing the Integral Time Absolute Error (ITAE), which is a desirable feature for Industrial IoT.

Toward the Design of an Efficient and Secure System Based on the Software-Defined Network Paradigm for Vehicular Networks

The advent of Intelligent Transport Systems (ITS) has led to the appearance of vehicles increasingly connected to their environment on global road networks. Due to the strict requirements for low latency and secure interactions in a vehicular environment, the proposal of new architectures is a crucial topic for discussion. This paper aims to develop a vehicular network using several access technologies based on SDN (Software Defined Network) paradigm, to take advantage of the capacities of the various access networks and provide flexibility in their control and management. Confidentiality, integrity, and authentication are essential services to prevent an adversary from compromising the security of vehicular networks. Therefore, good security and privacy management system is necessary to ensure this protection. We represent then a hybrid SDN-VANET architecture that can address all of the challenges we mentioned earlier. We are in the process of implementing a dynamic approach to optimize the positioning of controllers according to changes in network topology due to fluctuations in road traffic. We will also detail the topology estimation service based on machine learning techniques to provide network control functions with potential insight into the future state of the network, unlocking proactive and intelligent network control. We also provide a scheme that prevents and informs about basic and compound attacks and reacts to the privacy and security conditions of the vehicular network, managing the requirements of security management systems. The simulation results showed the effectiveness of the proposed schemes in terms of message loss rates, packet delivery rates (PDR), Round Trip Time (RTT), and delays. With used our scheme, the performance of the network is improved when SDN triggers the change of the RSU entity. Such as, we notice that the average RTT is lowered by 68 ms and that the PDR remains around 94%. We also notice with the integration of the security and privacy scheme (SPS) that the performance of the network is improved, the average RTT is reduced by 51 ms and the PDR persists around 99%.

Intelligent Congestion Control of 5G Traffic in SDN using Dual-Spike Neural Network

Software Defined Networking (SDN) with centralized control provides a global view and achieves efficient network resources management. However, using centralized controllers has several limitations related to scalability and performance, especially with the exponential growth of 5G communication. This paper proposes a novel traffic scheduling algorithm to avoid congestion in the control plane. The Packet-In messages received from different 5G devices are classified into two classes: critical and non-critical 5G communication by adopting Dual-Spike Neural Networks (DSNN) classifier and implementing it on a Virtualized Network Function (VNF). Dual spikes identify each class to increase the reliability of the classification. Different metrics have been adopted to evaluate the proposed classifier's effectiveness: accuracy, precision, recall, Matthews Correlation Coefficient (MCC), and F1-Score. Compared with a convolutional neural network (CNN), the simulation results confirmed that the DSNN model could enhance traffic classification accuracy by 5%. The efficiency of the priority model also has been demonstrated in terms of Round Trip Time (RTT).

ANALYSIS OF NETWORK TRAFFIC THREATS ACROSS OSI MODEL LAYERS FOR DYNAMIC RTO CALCULATION IN THE CONTEXT OF COMBATING DDoS ATTACKS

This document provides an examination of current threats to network security, viewed through the lens of network traffic analysis at various OSI model layers. It delves into the different forms of Distributed Denial of Service (DDoS) attacks and their ramifications on the Transmission Control Protocol (TCP), with a specific focus on a critical parameter - the Retransmission Timeout (RTO). The text also divulges fundamental algorithms and techniques for calculating RTO, encompassing adaptive methodologies that harness machine learning and artificial intelligence for optimizing the TCP/IP stack. In particular, it offers insights into the functioning of the RTO calculation algorithm, a pivotal element ensuring the reliability of data transmission via TCP. The document elaborates on how this algorithm dynamically adjusts the RTO value based on network conditions and measured Round Trip Time (RTT) values. Furthermore, it furnishes formulas for computing RTO with diverse parameters. Moreover, the document explores the potential of employing machine learning and data analysis methodologies to detect and preempt DDoS attacks. It elucidates how contemporary technologies empower the use of these approaches to minimize false positives in identifying malicious traffic packets, thereby enhancing the effectiveness of safeguarding information systems. Additionally, it provides an illustration of software and hardware tools employed for the practical implementation of these algorithms in devices facilitating data transmission via Ethernet connections. In summary, this work offers insights into contemporary challenges and issues in the realm of network security, especially in the context of the escalating frequency of DDoS attacks. This information proves valuable for students and professionals engaged in the study of network security and the development of measures to fortify networks and systems.

Efficient Acceleration of Deep Learning Inference on Resource-Constrained Edge Devices: A Review

Successful integration of deep neural networks (DNNs) or deep learning (DL) has resulted in breakthroughs in many areas. However, deploying these highly accurate models for data-driven, learned, automatic, and practical machine learning (ML) solutions to end-user applications remains challenging. DL algorithms are often computationally expensive, power-hungry, and require large memory to process complex and iterative operations of millions of parameters. Hence, training and inference of DL models are typically performed on high-performance computing (HPC) clusters in the cloud. Data transmission to the cloud results in high latency, round-trip delay, security and privacy concerns, and the inability of real-time decisions. Thus, processing on edge devices can significantly reduce cloud transmission cost. Edge devices are end devices closest to the user, such as mobile phones, cyber–physical systems (CPSs), wearables, the Internet of Things (IoT), embedded and autonomous systems, and intelligent sensors. These devices have limited memory, computing resources, and power-handling capability. Therefore, optimization techniques at both the hardware and software levels have been developed to handle the DL deployment efficiently on the edge. Understanding the existing research, challenges, and opportunities is fundamental to leveraging the next generation of edge devices with artificial intelligence (AI) capability. Mainly, four research directions have been pursued for efficient DL inference on edge devices: 1) novel DL architecture and algorithm design; 2) optimization of existing DL methods; 3) development of algorithm–hardware codesign; and 4) efficient accelerator design for DL deployment. This article focuses on surveying each of the four research directions, providing a comprehensive review of the state-of-the-art tools and techniques for efficient edge inference.

A Novel Approach to Network Forensic Analysis: Combining Packet Capture Data and Social Network Analysis

Log data from computers used for network forensic analysis is ineffective at identifying specific security threats. Log data limitations include the difficulty in reconstructing communication patterns between nodes and the inability to identify more advanced security threats. By combining traditional log data analysis methods with a more effective combination of approaches, a more comprehensive view of communication patterns can be achieved. This combined approach can then help identify potential security threats more effectively. It's difficult to determine the specific benefits of combining Packet Capture (PCAP) and Social Network Analysis (SNA) when performing forensics. This article proposes a new approach to forensic analysis that combines PCAP and social network analysis to overcome some of the limitations of traditional methods. The purpose of this discovery is to improve the accuracy of network forensic analysis by combining PCAP and social network analysis to provide a more comprehensive view of network communication patterns. Network forensics, which combines pcap analysis and social network analysis, provides more comprehensive results. PCAP analysis is used to analyze network traffic, conversation statistics, protocol distribution, packet content and round-trip times. Social network analysis maps communication patterns between nodes and identifies the most influential key players within the network. PCAP analysis efficiently captures and analyzes network packets, and SNA provides insight into relationships and communication patterns between devices on the network.

Impact of buffering time reduction on false congestion detection in TCP vegas over OBS networks

Since Transmission Control Protocol (TCP) is the dominant protocol on the Internet and optical burst switching (OBS) networks represent the optical transmission solution that can meet future high bandwidth requirements, TCP and OBS networks represent a possible combination for the next-generation Internet. However, a problem with this combination model is that operations at the OBS layer need to be controlled so that false congestion detection at the TCP layer is minimised. Considering the combination of TCP Vegas and OBS, congestion detection at the TCP layer is based on the round-trip time (RTT), where congestion is detected if the current RTT exceeds a given RTTmax. The extra delay may be due to some actual loss, but it may also be due to the extended execution time caused by some OBS layer operations. Therefore, reducing the time of operations at the OBS layer will reduce false congestion detection at the TCP layer. This paper investigates the impact of reducing the buffering time at the ingress node on the false congestion detection rate at the TCP layer. The reduction of the burst buffering time is accomplished by nesting the offset time in the assembly time. The simulation results and analysis show that the transmission efficiency in TCP Vegas over OBS networks is significantly improved; the throughput remains high, and the delivery success rate is significantly increased compared to the conventional buffering mode.