Real-time Packet Research Articles

A real-time network traffic classification system using cost-sensitive deep neural networks

Due to the continuous growth of Internet and online applications, network traffic classification is not only becoming one of the most crucial disciplines in network management but is also becoming quintessential for providing advanced tasks such as Quality of Service and network security. Moreover, even though many studies have been undertaken in recent years, real-time encrypted traffic classification continues to be an important challenge in the field of network traffic classification. Therefore, in this paper a real-time network traffic classification system is proposed together with five new models. The real-time classification system classified each incoming real-time packets into appropriate classes of interest and the five new models make use of a cost-sensitive learning strategy to deal with the unbalanced data problem during the training phase. The proposed models, which are called Cost-Sensitive Long-Short Term Memory (CSLSTM), Cost-Sensitive Gated Recurrent Unit (CSGRU), Cost-Sensitive Convolution Neural Network (CSCNN), CSNN with LSTM and CSCNN with GRU, can handle both traffic categorization and application identification. These proposed models were compared with prominent methods in this field and the proposed CSCNN was observed to outperform the researched deep learning models by at least 4% to 16% in correctly classifying packets from the ISCX VPN-nonVPN dataset.

Implementation and optimization of Burg’s method for real-time packet loss concealment in networked music performance applications

AbstractIn networked music performance (NMP) applications, which entail real-time audio streaming over the Internet, strict latency requirements are needed to ensure a realistic interaction between geographically dispersed musicians. Thus, NMP applications typically leverage uncompressed audio and unreliable transport protocols to avoid unnecessary processing and re-transmission delays. Given that no guarantee on packet delivery is offered, NMP applications must deal with late/lost audio packets to mitigate the impact of the resulting audio artifacts on the quality of the playback audio stream. This paper explores an audio packet loss concealment (PLC) technique based on autoregressive (AR) models. In particular, it investigates the algorithmic implementation of Burg’s method and the parameters configuration that offers the best trade-off between prediction error and computational time requirements. The purpose is to find the most suitable solution capable of running on a Raspberry Pi 4B within the real-time audio boundaries imposed by NMP applications. Additionally, we analyze the computational time required to fit the model and predict future samples by considering six implementations and various compilation flags. Results confirm that AR models can predict future audio samples more accurately than traditional PLC approaches, which consist of filling audio gaps with silence or repeating the last received audio segment. Furthermore, results demonstrate the effectiveness of the proposed solution in meeting the strict latency requirements when deployed on a Raspberry Pi 4B.

Parameterization and Performance Analysis of a Scalable, near Real-Time Packet Capturing Platform

The rapid evolution of technology has fostered an exponential rise in the number of individuals and devices interconnected via the Internet. This interconnectedness has prompted companies to expand their computing and communication infrastructures significantly to accommodate the escalating demands. However, this proliferation of connectivity has also opened new avenues for cyber threats, emphasizing the critical need for Intrusion Detection Systems (IDSs) to adapt and operate efficiently in this evolving landscape. In response, companies are increasingly seeking IDSs characterized by horizontal, modular, and elastic attributes, capable of dynamically scaling with the fluctuating volume of network data flows deemed essential for effective monitoring and threat detection. Yet, the task extends beyond mere data capture and storage; robust IDSs must integrate sophisticated components for data analysis and anomaly detection, ideally functioning in real-time or near real-time. While Machine Learning (ML) techniques present promising avenues for detecting and mitigating malicious activities, their efficacy hinges on the availability of high-quality training datasets, which in turn poses a significant challenge. This paper proposes a comprehensive solution in the form of an architecture and reference implementation for (near) real-time capture, storage, and analysis of network data within a 1 Gbps network environment. Performance benchmarks provided offer valuable insights for prototype optimization, demonstrating the capability of the proposed IDS architecture to meet objectives even under realistic operational scenarios.

OPTIMIZING EFFICIENT AND RELIABLE PACKET TRANSMISSION OVER WIRELESS NETWORKS

Wireless Networks are becoming an increasingly important technology that is bringing the world closer together. In this type of network environment there could be more chances of attacks. The packets cannot be easily transferred over the network. It affects network performance degrade. While eavesdropping and message injection can be prevented using cryptographic methods, jamming attacks are much harder to counter. They have been shown to actualize severe Denial-of- Service (DoS) attacks against networks. In Simplest form adversary blocks the packets that are transmitted over wireless network. Typically, jamming attacks has been considered under an external threat model, in which the jammer is not part of the network. To overcome the above problem of network traffic and performance in this paper we have considered a packet hiding methods that can be securely transmit packets over the network. We are addressing the problem of jamming attacks under internal threat model and two schemes are proposed that prevent real-time packet classification of packets by combining hiding scheme based on cryptographic primitives. Keywords: Selective Jamming, Denial-of-Service, Wireless Networks, Packet Classification.

Bandwidth allocation of URLLC for real-time packet traffic in B5G: A Deep-RL framework

By considering the limited energy of Internet of Things (IoT) devices. We take the resource allocation to guarantee the stringent Quality of Service (QoS) depending on the joint optimization of power control and finite blocklength of channel. To achieve large volumes of arrival rates, we propose Adversarial Training based Generative Adversarial Networks (AT-GANs), which utilize a significant number of extreme events to provide high reliability and adjust real data in real-time. Simulation results show that Deep-Reinforcement Learning (Deep-RL) for AT-GAN could eliminate the transient training time. As a result, the AT-GAN keeps the reliability higher than 99.9999%.

Construction of a Digital Shared Management Platform for Enterprise Finance in the Information Age

Abstract Based on the construction objectives of the enterprise financial sharing management platform, this paper focuses on the design concept and implementation process of each system module based on the overall system design under the framework of the financial sharing platform. The BP neural network algorithm is used to evaluate the financial information risk, the particle swarm optimization algorithm is used to optimize the database, the big data visualization technology is used for digitizing the financial information, and the metadata sharing technology is used to transfer the financial sharing data, to complete the financial information sharing. Through testing the platform, it is concluded that the real-time packet loss rate of financial sharing data is kept below 20%, and the average response time of multi-user clicks is within 2S. Therefore, the sharing function, the function of each module, and the performance of the platform in this paper can meet the needs of enterprises.

Real-Time Guarantees in Routerless Networks-on-Chip

This article considers the use of routerless networks-on-chip as an alternative on-chip interconnect for multi-processor systems requiring hard real-time guarantees for inter-processor communication. It presents a novel analytical framework that can provide latency upper bounds to real-time packet flows sent over routerless networks-on-chip, and it uses that framework to evaluate the ability of such networks to provide real-time guarantees. Extensive comparative analysis is provided, considering different architectures for routerless networks and a state-of-the-art wormhole network based on priority-preemptive routers as a baseline.

Cross-Layer Based QoS Aware Load-balancing Multi-Path Routing Protocol over Wireless Multimedia Sensor Networks

Abstract: Currently, the utilization of WMSNs in different real-time and non-real-time applications requires an excessive amount of bandwidth for reliable data delivery. The unique features of WMSNs are significantly challenging in satisfying the QoS requirements in such application-specific environments and balancing the traffic load among the devices. The provision of reliable multipath routing is a cornerstone in fulfilling the QoS requirements of WMSNs. Selecting multiple optimal paths between a source and destination based on peculiar routing metrics enhances the performance of QoS routing. Generally, routing protocols exploit several routing metrics, such as delay, remaining energy of nodes, hop count, available bandwidth, and packet loss rate in path selection to attain high reliability in data delivery. Many existing routing protocols only consider the network layer parameters, whereas it lacks focus on the data link and physical layer parameters, which creates a severe impact on the degradation of QoS. In addition to that, varying bandwidth channels create interference in multimedia data delivery and degrade the network performance. Designing a multipath routing protocol by considering cross-layer parameters offer a promising solution to optimize the WMSN performance. In cross-layer design, diverse protocol layers support the routing decisions adaptively by perceiving the dynamic characteristics of the wireless medium, resulting in fair use of scarce resources with high QoS. A Cross-Layer Based QoS Aware Load-Balancing Multipath Routing Protocol over Wireless Multimedia Sensor Networks was the goal of the study's five design objectives. The study and analysis of QoS and cross-layer-based routing algorithms for WMSNs was the initial goal. Secondly, a Deep Learning prioritization-based packet classifier to divide traffic according to priority. To ensure fair resource consumption and distribution of multimedia traffic, the third goal was to design and create a cross-layer optimizer model for optimal multiple disjoint route selection using machine learning techniques. The development of a cutting-edge channel-scheduling algorithm was goal four. It was designed to efficiently assign low-interference channels to communication devices in order to lower the packet drop rate in real-time packet delivery. Last but not least, a security method for Wireless Multimedia Sensor Networks' Cross-Layer based multipath routing protocol.

Evaluating the Performance of Proposed Switched Beam Antenna Systems in Dynamic V2V Communication Networks.

This paper develops a novel approach for reliable vehicle-to-vehicle (V2V) communication in various environments. A switched beam antenna is deployed at the transmitting and receiving points, with a beam management system that concentrates the power in each beam using a low-computation algorithm and a potential mathematical model. The algorithm is designed to be flexible for various environments faced by vehicles. Additionally, an anti-failure system is proposed in case the intelligent transportation system (ITS) system fails to retrieve real-time Packet Delivery Ratio (PDR) values related to traffic density. Performance metrics include the time to collision in seconds, the bit error rate (BER), the packet error rate (PER), the average throughput (Mbps), the beam selection probability, and computational complexity factors. The proposed system is compared with traditional systems. Extensive experiments, simulations, and comparisons show that the proposed approach is excellent and reliable for vehicular systems. The proposed study demonstrates an average throughput of 1.7 Mbps, surpassing conventional methods' typical throughput of 1.35 Mbps. Moreover, the bit error rate (BER) of the proposed study is reduced by a factor of 0.1. Additionally, the proposed framework achieves a beam power efficiency of touching to 100% at computational factor of 34. These metrics indicate that the proposed method is both efficient and sufficiently robust.

Framework for identifying network attacks through packet inspection using machine learning

Abstract In every network, traffic anomaly detection system is an essential field of study. In the communication system, there are various protocols and intrusions. It is still a testing area to find high precision to boost the correct distribution ratio. Many authors have worked on various algorithms such as simple classification, K-Means, Genetic Algorithm, and Support Vector Machine approaches, and they presented the efficiency and accuracy of these algorithms. In this article, we have proposed a feature extraction technique known as “k-means clustering,” which has its roots in signal processing and is employed to divide a set of n observations into k clusters, each of which has its origin from the observation with the closest mean. K-Means method is applied in this study to investigate the stream and its implementation and applications using Python and the dataset on the KDDcup99. The effectiveness of the outcome indicates the planned work’s efficiency in relation to other widely available alternatives. Apart from the applied method, a web-based framework is designed, which can inspect an actual network traffic packet for identifying network attacks. Instead of using a static file for testing the network attack, a web page-based solution uses database to collect and test the information. Real-time packet inspection is provided in the proposed work for identifying new attacks.

ESnet/JLab FPGA Accelerated Transport

To increase the science rate for high data rates/volumes, Thomas Jefferson National Accelerator Facility (JLab) has partnered with Energy Sciences Network (ESnet) to define an edge to compute cluster traffic shaping / steering transport capability featuring data event aware network shaping and forwarding. The keystone of this ESnet+JLab FPGA Accelerated Transport (EJFAT) is the joint development of a dynamic compute work Load Balancer (LB) of UDP streamed data. The LB is a suite consisting of a Field Programmable Gate Array (FPGA) executing the dynamically configurable, low fixed latency LB <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">data plane</i> featuring real-time packet redirection and high throughput, and a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">control plane</i> running on the FPGA host computer that monitors network and compute farm telemetry in order to make dynamic load-balancing decisions for destination compute host redirection / load balancing. The LB provides for three-tier horizontal scaling across LB suites, cluster compute hosts, and CPUs within a host. The LB effectively provides seamless integration of edge / cluster computing to support direct experimental data processing for immediate use by JLab science programs and others such as the Electron-Ion Collider (EIC) as well as data centers of the future requiring high throughput and low latency for both time-critical (e.g., data acquisition systems) and data-driven (data center) workflows.

Impact of fluctuating initial conditions on bottomonium suppression in 5.02 TeV heavy-ion collisions

We compute bottomonium suppression and elliptic flow within the pNRQCD effective field theory using an open quantum systems approach. For the hydrodynamical background, we use 2+1D MUSIC second-order viscous hydrodynamics with IP-Glasma initial conditions and evolve bottom/antibottom quantum wave packets in real time in these backgrounds. We find that the impact of fluctuating initial conditions is small when compared to results obtained using smooth initial conditions. Including the effect of fluctuating initial conditions, we find that the Upsilon(1S) integrated elliptic flow is v_2[1S] = 0.005 +/- 0.002 +/- 0.001, with the first and second variations corresponding to statistical and systematic theoretical uncertainties, respectively.

LossDetection: Real-Time Packet Loss Monitoring System for Sampled Traffic Data

Packet loss is common in networks, which leads to network quality of service degradation. Packet loss is an essential and concerning symptom when the quality of service is degraded. Therefore, real-time passive packet loss detection is conducive to estimating network services. Existing passive packet loss detection methods mainly study the packet loss for TCP using header information from full traffic. However, it cannot infer packet loss status for UDP due to its limited header information and is too costly to perform full acquisition in real networks. To address these problems, we propose a framework called LossDetection based on packet sampling and Feature-Sketch to detect packet loss in real time for both TCP and UDP. The result shows that our methodology can detect packet loss with an accuracy of 98%-100% at a sampling rate of 1/16. Furthermore, our extensive evaluation demonstrates that LossDetection is easy to implement in a software router and achieves low memory and detection latency while providing real-time information about packet loss.

Deep Learning-based Prioritized Packet Classification and Optimal Route Selection in Wireless Multimedia Sensor Networks

Wireless multimedia sensor networks suffer network congestion due to increased traffic leading to data packet loss and increased energy consumption. Thus, it is essential to have a mechanism for priority-based packet classification to improve the performance and decrease the delay in real-time packets. This paper introduces priority-based packet classification using a word embedding mechanism to extract packet semantics and classification using the long short-term memory model that works based on the different characteristics of the packet header to decide the priority of each packet. Furthermore presents a hybrid meta-heuristic algorithm termed butterfly-based rider optimization algorithm, which considers the packet priority and other parameters in optimal route selection. This hybrid meta-heuristic algorithm is developed by merging rider optimization algorithm, and butterfly optimization algorithm. Finally, the experimental results validate the analytical result as well as the performance comparison between the proposed optimal route selection model and the conventional models.

RuleOut Forwarding Anomalies for SDN

Reliable Software-Defined Networking (SDN) should mitigate forwarding anomalies due to cross-plane rule inconsistencies. Most existing countermeasures either inject probe packets to infer forwarding correctness or collect packet traces to detect forwarding anomalies. They, however, cannot detect or filter forwarding anomalies for production packets in real time. In this paper, we propose RuleOut as the first attempt to automatically throttle SDN forwarding anomalies. It disambiguates dependent rules via augmenting their matching fields with unique tags. Leveraging source routing, we further bind each packet with the tag sequence corresponding to rules the packet should match. RuleOut thus renders each packet to match at most one rule on each switch. This completely addresses the root cause of forwarding ambiguity. To implement RuleOut, we develop a non-overlapping rule dependency graph, a series of algorithms for incremental rule update and tag generation upon it, and various optimization techniques toward scalability and efficiency. We prototype RuleOut on the Ryu controller and Open vSwitch and evaluate its performance over public rule sets such as Stanford, Internet2, and Airtel1. RuleOut can use tags of only several bits long to disambiguate thousands to millions of rules and generate tags fairly fast within a few milliseconds.

An Adaptive Confidentiality Security Service Enhancement Protocol Using Image-Based Key Generator for Multi-Agent Ethernet Packet Switched Networks

Conventional network Security Protocols provide different types of security services that establish secure connections between network nodes making them robust against different security threats. The security establishment process depends on sharing security keys between connected devices. The level of security provided by such security protocols is static and irrelevant to the status of the network. In this paper, we propose an adaptive security-aware lottery-EDF scheduler for limited resources Ethernet packet switched networks using GAIA multi-agent approach. The proposed system adopts a periodic non-shared cryptographic key-generation process of different sizes (i.e., 64 bits, 128 bits, 160 bits, and 512 bits) using the textural features of digital images stored on a central image-server. The proposed system deploys a security enhancement unit that adaptively enhances the confidentiality security-service level for the real-time packets according to feedback from a congestion control scheme that is based on network resource estimation technique, hence making the network robust against the spoofing security threat while preserving the overall network performance metrics (NPMs). By simulating a real-time Ethernet packet switched network, extensive simulation results show the efficiency of the adaptive proposed protocol over the static-security and IPSec protocols in guaranteeing the QoS requirements for the real-time flows in terms of destination’s buffer consumption and average total packet delays.

A high-performance dual classifier based packet classification engine on FPGA

Security is one of the key elements of real-time communication systems, especially with the developments in digital communication systems. One of the most crucial networking activities that guards against hacking and unintentional consequences on connected devices is packet classification. The router's and other linked devices' sluggish processing speeds are a problem for implementing real-time packet classification. This paper proposes a novel parallel architecture for packet classification in routers. The design that has been suggested uses the parallelism that modern hardware technologies offer to categorize several packets at once to increase throughput, lower energy usage, and drastically lower latency. The suggested design supports multiple data reads by utilizing Dual Port RAM, or DPRAM. They have low latency performance and allow for fast rate packet lookup, making them the preferable choice for stage memory. The DPRAM provides the data to the packet scheduling module. The packets are received concurrently by the two subsequent packet splitter units. The throughput of packet classification increases as a result. In this paper, a parallel packet categorization architecture using FPGA is explored for maximizing throughput and maintain allowable latency while preserving low energy consumption. The classifier has the lowest energy efficiency of 7.2nJ and less latency of 58 ns. Its throughput may reach 668 MPPS (million packets per second).

PCMIgr: a fast packet classification method based on information gain ratio

To solve the problem of ambiguous attribute selection in existing decision tree classification algorithms, a decision tree construction method based on information entropy, PCMIgr, is proposed. PCMIgr is a heuristic method based on greedy strategy. At each decision tree node, when it is necessary to select classification attributes for division, the attribute with the highest information gain ratio is selected. The main innovation of this method is that the attribute selection in the traditional classification method based on decision tree is optimized, and the classification efficiency of the constructed decision tree is improved compared with that before optimization. At the same time, the decision tree ensures that each leaf node is only associated with one rule, which avoids the common problem of "rule replication" in the process of traditional decision tree construction, and effectively saves memory and calculation time. The experimental results show that the application of this method to the construction of classification decision tree can further improve the efficiency of packet classification method based on decision tree, and can be applied to high-speed real-time packet classification.

Design, implementation, and performance evaluation of an earliest-deadline-first packet scheduling scheme in P4 hardware switches

Real-time applications such as video conferencing, virtual reality (VR), high-speed train controls, nuclear plant controls, etc. impose stringent limitations on the maximum end-to-end latency that their packets can tolerate. In such applications, if a packet arrives at its destination node but misses its deadline, it will be useless. On a network, a packet often needs to pass multiple switches on its way from its source node to its destination node. In some of these switches, the packet may experience large queuing delays and the accumulation of these delays may cause the packet to miss its deadline at its destination node. In this work, we design and implement an earliest-deadline-first (EDF) packet scheduling scheme in P4 hardware switches to overcome this problem. We implement two versions of our scheme that use two queues and six queues respectively for real-time packets to study how the number of queues may affect the performance of our scheme. We evaluate the performance of our scheme under different levels of congestion conditions. Experimental results show that, compared with the baseline scheme that dispatches all real-time packets to the same high-priority queue, our scheme greatly reduces the deadline miss rate of real-time packets when the real-time traffic ratio in the network is between 40% and 80%.

Network-on-Chip Irregular Topology Optimization for Real-Time and Non-Real-Time Applications.

Network-on-Chip is a good approach to working on intra-chip communication. Networks with irregular topologies may be better suited for specific applications because of their architectural nature. A good design space exploration can help the design of the network to obtain more optimized topologies. This paper proposes a way of optimizing networks with irregular topologies through the use of a genetic algorithm. The network proposed here has heterogeneous routers that aim to optimize the network and support applications with real-time tasks. The goal is to find networks that are optimized for average latency and percentage of real-time packets delivered within the deadline. The results show that we have been able to find networks that can deliver all the real-time packets, obtain acceptable latency values, and shrink the chip area.