Improve TCP Performance Research Articles

Improving TCP performance over a common IoT scenario using the Early Window Tailoring method

The Internet of Things (IoT) is already a reality in homes, transforming the home network into a complex composition of different equipment where each one has its own attributes in terms of TCP protocol variants, transmission speed, reliability and delay. In home network scenarios where IoT devices are used, it is common to have a 802.11 environment with several devices sharing the same Access Point (AP). When TCP is used in this environment its performance is affected by many things, such as wireless link losses, queue saturation on access points, and fairness. To improve the TCP performance in these scenarios, this paper presents a new version of the Early Window Tailoring (EWT) approach. The EWT-IoT do not require any modification in the TCP, and it does not require that all routers in the path use the same Active Queue Management (AQM), the use in the AP is sufficient. To evaluate its performance, comparison tests against drop-tail, RED, ARED and EWA were performed. The EWT-IoT proved to be effective in congestion control, achieving greater performance than the other approaches, reducing the transfer time, having a superior goodput, maintaining a satisfactory fairness, and presenting no losses. Also, the EWT-IoT allowed the existence of a number of flows on average 65.2% better than its best competitor and 71.3% better when no AQM scheme was used.

R-AQM: Reverse ACK Active Queue Management in Multitenant Data Centers

TCP incast has become a practical problem for high-bandwidth, low-latency transmissions, resulting in throughput degradation of up to 90% and delays of hundreds of milliseconds, severely impacting application performance. However, in virtualized multi-tenant data centers, host-based advancements in the TCP stack are hard to deploy from the operators’ perspective. Operators only provide infrastructure in the form of virtual machines, in which only tenants can directly modify the end-host TCP stack. In this paper, we present R-AQM, a switch-powered reverse ACK active queue management (R-AQM) mechanism for enhancing ACK-clocking effects through assisting legacy TCP. Specifically, R-AQM proactively intercepts ACKs and paces the ACK-clocked in-flight data packets, preventing TCP from suffering incast collapse. We implement and evaluate R-AQM in NS-3 simulation and NetFPGA-based hardware switch. Both simulation and testbed results show that R-AQM greatly improves TCP performance under heavy incast workloads by significantly lowering packet loss rate, reducing retransmission timeouts, and supporting 16 times (i.e., 60 to 1000) more senders. Meanwhile, the forward queuing delays are also reduced by 4.6 times.

Coefficient of Restitution based Cross Layer Interference Aware Routing Protocol in Wireless Mesh Networks

In Multi-Radio Multi-Channel (MRMC) Wireless Mesh Networks (WMN), Partially Overlapped Channels (POC) has been used to increase the parallel transmission. But adjacent channel interference is very severe in MRMC environment; it decreases the network throughput very badly. In this paper, we propose a Coefficient of Restitution based cross layer interference aware routing protocol (CoRCiaR) to improve TCP performance in Wireless Mesh Networks. This approach comprises of two-steps: Initially, the interference detection algorithm is developed at MAC layer by enhancing the RTS/CTS method. Based on the channel interference, congestion is identified by Round Trip Time (RTT) measurements, and subsequently the route discovery module selects the alternative path to send the data packet. The packets are transmitted to the congestion free path seamlessly by the source. The performance of the proposed CoRCiaR protocol is measured by Coefficient of Restitution (COR) parameter. The impact of the rerouting is experienced on the network throughput performance. The simulation results show that the proposed cross layer interference aware dynamic routing enhances the TCP performance on WMN.

Improving TCP Performance over Mobile Ad Hoc Networks with an Adaptive Backoff and Response approach (Dept.E)

window.dataLayer = window.dataLayer || []; function gtag(){dataLayer.push(arguments);} gtag("js", new Date()); gtag("config", "UA-72322659-1");

Improving TCP Performance with Intelligent Loss Differentiation for Mobile Adhoc Networks

Improving TCP Performance with Intelligent Loss Differentiation for Mobile Adhoc Networks

Improving TCP Performance Over WiFi for Internet of Vehicles: A Federated Learning Approach

We propose a novel communication efficient and privacy preserving federated learning framework for enhancing the performance of Internet of Vehicles (IoV), wherein on-vehicle learning models are trained by exchanging inputs, outputs and their learning parameters locally. Moreover, we use analytic modeling as a tool for reasoning and developing the required IoV scenario and stabilize their data flow dynamics by considering TCP CUBIC streams over WiFi networks to prove our idea.

Performance Analysis of TCP New Reno Over Satellite DVB-RCS2 Random Access Links

In this paper, the performance evaluation of TCP New Reno over satellite DVB-RCS2 links with random access is analyzed. Based on the satellite system consisting of a large number of source-destination TCP New Reno pairs, a Markov chain model for a TCP connection is conducted. According to the principles of TCP New Reno, two one-step transition probabilities for the value of congestion window in different data transmission phases (e.g. the slow start phase and congestion avoidance phase) are proposed. In order to analysis the throughput of a TCP connection, according to the packets lost events, three conditions (no packet lost, packets lost in the congestion avoidance phase, and packets lost in the slow start phase), are theoretically analyzed. We analytically derive the upper and lower throughput bounds, simulations and analysis are integrated together to indicate the effective ness and correctness of theoretical analysis. Moreover, we can clearly determine which factor plays role for the performance of TCP New Reno. The analysis not only includes the main algorithms of TCP protocol, but also is developed for general random access. Therefore, the Markov model and the theoretical analysis can be applied to many erasure channel, and the conclusions are extremely critical for the research of improving TCP performance over links with random access.

TCP Stalls at the Server Side: Measurement and Mitigation

TCP is an important factor affecting user-perceived performance of Internet applications. Diagnosing the causes behind TCP performance issues in the wild is essential for better understanding the current shortcomings in TCP. This paper presents a TCP flow performance analysis framework that classifies causes of TCP stalls. The framework forms the basis of a tool that we use to analyze packet-level traces of three services (cloud storage, software download, and web search) deployed by a popular service provider. We find that as many as 20% of the flows are stalled for half of their lifetime. Network-related causes, especially timeout retransmissions, dominate the stalls. A breakdown of the causes for timeout retransmission stalls reveals that double retransmission and tail retransmission are among the top contributors. The importance of these causes depends however on the specific service. Based on these observations, we propose smart-retransmission time out (S-RTO), a mechanism that mitigates timeout retransmission stalls through careful and gentle aggression for retransmission. S-RTO is evaluated in a controlled network and also in a production network. The results consistently show that it is effective at improving TCP performance, especially for short flows.

무선 환경에서 TCP 성능 개선을 위한 머신러닝 기반 LDA 혼잡제어 설계

TCP는 신뢰성을 기반으로 한 대표적인 전송계층 프로토콜이다. 하지만 TCP는 과거 80년대에 설계된 프로토콜이기 때문에 유선환경에 적합하게 설계되어있다. 이러한 이유로 현재 TCP에서 사용하는 손실 기반의 혼잡제어는 어떠한 손실이라도 혼잡으로 판단하여 동작하기 때문에 무선 채널 에러가 자주 발생하는 무선 환경에서 성능 저하가 발생하는 문제가 있다. 본 논문에서는 이와 같은 문제를 해결하기 위해 머신러닝 기반의 LDA(Loss Discriminiation Algorithm)을 설계하였으며 이를 적용한 MLLDA(Machine Learning based-LDA) 혼잡제어를 제안하였다. MLLDA 혼잡제어는 머신러닝 기법 중에서 MLP를 사용하여 혼잡으로 인한 패킷 손실과 무선 채널 환경에 의한 손실을 구분하도록 학습시켰다. 그리고 학습 결과를 바탕으로 혼잡제어를 수행하도록 설계하였다. 제안한 혼잡제어 검증을 위해 리눅스커널에 제안한 알고리즘을 구현하고 패킷 손실이 발생하는 테스트베드를 구성하여 기존 혼잡제어와 비교분석 하였다. 실험결과 제안한 혼잡제어가 패킷 손실이 발생하는 환경에서 패킷 손실구분 정확도가 98%로 나타났고, 기존 혼잡제어보다 패킷 손실 구분 정확도와 평균 전송속도 성능이 크게 개선됨을 보였다.

Solution for packet reordering in manet’s using cross layer based routing protocol

TCP allows sharing the information about network status using Cross Layer mechanism for the different layers. The Cross Layer approach in wireless MANETs is provided for improving TCP performance. The Optimal Path or Route Selection is an important aspect for increasing energy efficiency and lifetime of the network. A proposal for Trust Aware Routing protocol is described for selecting an optimal route in MANETs. Direct and Indirect trust values are used for estimating the trust value of every node in the network. The route cost amount is estimated and the best path with minimum cost value is chosen for the network. The data packet is transmitted on the optimal path, and the optimal path is selected on the minimum cost value of the network. There is a possibility of the packets being dropped due to congestion and / or mobility and there are chances of the packets being reordered during the transmission process. Reorder Identification mechanism is the procedure used for dropped and / or reordered packets. The simulation results of the proposed work are compared with the existing work and it shows more effective in terms of energy efficiency and network lifetime.

Convex Optimized Lagrange Multiplier Based Algebraic Congestion Likelihood for Improved TCP Performance in MANET

Convex Optimized Lagrange Multiplier Based Algebraic Congestion Likelihood for Improved TCP Performance in MANET

Wireless Networking Performance in IoT Using Adaptive Contention Window

Internet of Things (IoT) network contains heterogeneous resource‐constrained computing devices which has its unique reputation in IoT environments. In spite of its distinctiveness, the network performance deteriorates by the distributed contention of the nodes for the shared wireless medium in IoT. In IoT network, the Medium Access Control (MAC) layer contention impacts the level of congestion at the transport layer. Further, the increasing node contention at the MAC layer increases link layer frame drops resulting in timeouts at the transport layer segments and the performance of TCP degrades. In addition to that, the expiration of maximum retransmission attempts and the high contentions drive the MAC retransmissions and the associated overheads to reduce the link level throughput and the packet delivery ratio. In order to deal with aforementioned problems, the Adaptive Contention Window (ACW) is proposed, which aims to reduce the MAC overhead and retransmissions by determining active queue size at the contending nodes and the energy level of the nodes to improve TCP performance. Further, the MAC contention window is adjusted according to the node’s active queue size and the residual energy and TCP congestion window is dynamically adjusted based on the MAC contention window. Hence, by adjusting the MAC Adaptive Contention Window, the proposed model effectively distributes the access to medium and assures improved network throughput. Finally, the simulation study implemented through ns‐2 is compared with an existing methodology such as Cross‐Layer Congestion Control and dynamic window adaptation (CC‐BADWA); the proposed model enhances the network throughput with the minimal collisions.

Performance evaluation of TCP over software-defined optical burst-switched data centre network

In this paper, we consider the performance of TCP when used in data centre networks (DCNs) featuring optical burst switching (OBS) using two-way reservation. The two-way reservation is not suitable in wide-area OBS networks due to high bandwidth-delay product (BDP). The burst loss using traditional methods of one-way reservation can be mistakenly interpreted by the TCP layer as congestion instead of contention in OBS network, leading to serious degradation of the TCP performance. The reduced BDP in DCNs allows the use of two-way reservation that results in zero burst loss. The modelled architecture features fast optical switches in a single hop topology. We apply different workloads with various burst assembly parameters to evaluate the performance of TCP. Our results show significant improvement in TCP performance as compared to traditional methods of OBS as well as to a conventional electronic packet switching DCN.

Investigation of Smart Meter Data Reporting Strategies for Optimized Performance in Smart Grid AMI Networks

Designing efficient and reliable wireless mesh-based advanced metering infrastructure (AMI) networks is challenging. In AMI networks, fine-grained regular data collections from smart meters (SMs) create a lot of traffic and interference. The location of the gateway that collects data from SMs may also add to this interference by impacting the length of routes. Furthermore, TCP-like protocols that are employed for reliability may bring additional overhead. Therefore, it is critical to pick the suitable data collection strategy and gateway location to meet some smart grid performance requirements. In this paper, we proposed three novel data collection mechanisms to set the periodic reporting time of each SM to improve TCP performance in IEEE 802.11s-based wireless mesh AMI networks. The first idea was based on the nature of IEEE 802.11s routing protocol. Each SM is assigned a reporting time based on its location in the spanning tree network. The second idea was inspired by the time division multiple access methods where each meter is given a separate slot. The third idea was based on both previous ideas and clustering to increase the number of meters that can send at the same slot. For the gateway location, we also proposed a novel mechanism based on ${p}$ -center facility problem to minimize data delivery delay. The simulation results indicate that the packet delay can be improved significantly without any negative impact on the other performance metrics.

LIFT: a Local IPSec-aware Freezing Protocol to improve TCP Performance in Satellite Networks

In this paper a protocol, local to the satellite link, is defined in order to boost TCP performance in mobile integrated wired-satellite Internet. It has been conceived to help to overcome the well known retransmissions competition problem that arises when a satellite reliable link layer protocol is used to face satellite link errors. This protocol, called Local IPSec-aware Freezing proTocol (LIFT), has been designed to allow the satellite gateway, even in the presence of communications secured by IPSec, to freeze the TCP sender when it perceives a possible delay due to satellite channel conditions. The effectiveness of LIFT has been evaluated, using the ns-2 tool, in terms of Web page download mean time for a satellite mobile host. Simulation results have shown that the adoption of LIFT protocol provides substantial improvements in TCP performance.

A fuzzy-based mechanism for delay window adjustment to improve TCP performance in wireless multi-hop networks

A fuzzy-based mechanism for delay window adjustment to improve TCP performance in wireless multi-hop networks

A fuzzy-based mechanism for delay window adjustment to improve TCP performance in wireless multi-hop networks

TCP protocol was initially designed for wired networks to guarantee reliable data transfer. In wired networks, TCP assumes congestion if data packets are dropped. However, this assumption does not hold when the end-to-end path includes wireless links, where, high bit error rate (BER), shared medium and dynamic nature of the wireless channel may cause packet loss. This results in a severe degrade in performance of standard TCP in wireless networks. On the other hand, TCP cannot be changed fundamentally due to the large base of installation in the wired network. Delayed-ACK generation at the TCP receiver has been proposed to mitigate collision rate within wireless network. However, choosing appropriate delayed window size is an important issue. In this article, we propose a new delayed-ACK approach based on a fuzzy controller to improve TCP performance in wireless networks. The major advantage of our mechanism is reducing the overhead and the collision rate in multi-hop wireless networks. We also compare our proposal with some of well-known studies. Simulations results confirm good performance of our mechanism.

An Approach Using Dynamic Deferred Acknowledgement to Improve TCP Performance in Multi-Hop Wireless Networks

An Approach Using Dynamic Deferred Acknowledgement to Improve TCP Performance in Multi-Hop Wireless Networks

An Efficient Explicit Error Notification with Adaptive Packet Size Mechanism to Improve TCP Performance in LEO Network

An Efficient Explicit Error Notification with Adaptive Packet Size Mechanism to Improve TCP Performance in LEO Network

A new explicit congestion notification scheme for satellite IP networks

Longer transmission delay and higher link errors affect data communication performance over satellite links. The problem gets compounded in case a congestion occurs in the corresponding data path. Taking preventive measures before a congestion actually occurs can help in avoiding such situations. Irrespective of the network characteristics, Explicit Congestion Notification (ECN) either through ICMP messages or through marking packets aims to achieve this objective. However, use of ICMP messages for explicit congestion notification leads to vulnerability to attacks by malicious hosts. On the other hand, use of conventional marking schemes that mark data packets in the forward direction can be slow as they have to traverse the satellite links. In this manuscript, we propose a new mechanism of ECN for an even faster notification of an incipient congestion over satellite IP networks. Our proposal, Mark-Reverse ECN, marks packets corresponding to acknowledgement (ACK) segments in the reverse direction to the sender. This leads to a reduction of the latency between detection and actual notification. As such, our mechanism leads to a more accurate and effective approach to congestion control. In addition, our proposed mechanism is free from vulnerability to malicious attacks. We also develop a new analytical model for Mark-Reverse ECN on TCP Reno for performance prediction and validation of simulation outcomes. Simulation results show that our proposed mechanism (i) leads to up to 25% improvement of TCP performance, (ii) reduces packet loss at the congested router by down to about 0% and (iii) helps to improve the performance of conventional TCP when deployed together.