iTCP: an intelligent TCP with neural network based end-to-end congestion control for ad-hoc multi-hop wireless mesh networks

Islam, Ariful; Raghunathan, Vijay

doi:10.1007/s11276-014-0799-6

Cited by 25 publications

(1 citation statement)

References 83 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Para los usuarios, este proceso de interconexión inalámbrica es totalmente transparente, es decir, los usuarios no perciben que la iWMN actúa como intermediaria entre dos tecnologías de red, es decir, la red IEEE 802.11 Wi-Fi y la red IEEE 802.3 Ethernet. A pesar de las características destacadas de las iWMNs, se ha demostrado que sus prestaciones pueden degradarse por numerosos factores; tales como: la alta densidad de usuarios en la red [2], la distribución sesgada del tráfico [3], el incremento de la contienda por el acceso al medio [4], así como por el pobre desempeño que tiene el protocolo TCP en redes inalámbricas [5,6]. Estos hallazgos ponen en duda los beneficios que pudieran obtenerse al utilizar esta tecnología.…”

Section: Capítulounclassified

Codificación adaptativa de red para sistemas inalámbricos IEEE 802.11s en modo infraestructura

Arellano

View full text Add to dashboard Cite

IEEE 802.11S INFRASTRUCTURE WIRELESS MESH NETWORKS (commonly known as iWMNs) are integrated by static wireless nodes capable of working in coordination to route data packets. In this way, the nodes collaborate to exchange information with each other. In addition, iWMNs can be interconnected with other network technologies and, in this way, help to wirelessly extend the coverage of these networks; for example, iWMNs are used today to extend the coverage of cellular or wired networks. Thanks to this feature, and also to their low infrastructure cost, iWMNs networks are considered today as an excellent option to offer wireless Internet connectivity services in geographical areas where the use of other network technologies is unfeasible. Despite the promising features of iWMNs, there are studies and results that cast doubt on their performance, since it has been documented that the performance of these networks can be affected by numerous factors; such as the use of TCP to transport information in wireless environments, the transmission errors in the wireless medium, as well as the access contention between network users. All these factors can degrade the performance of iWMNs and, consequently, affect the quality of the experience for the users. In this doctoral thesis, some of these performance problems are addressed through the technique called adaptive network coding. With this technique, the nodes of an iWMN are allowed to combine various data packets and thus build an encoded packet; this packet contains the information from the original packets, requiring only one wireless transmission to transport the original information, reducing the use of the wireless medium and, thereby, increasing the capacity of the network. The proposed technique also seeks to adapt the coding process to the traffic conditions in the network through the dynamic adjustment of the waiting time of the packets in a node before they can be combined. This proposal aims to substantially improve the performance of iWMNs, solving some problems that affect them. The evaluation of the proposal is carried out through simulations and numerical evaluations. After a detailed analysis of the results, we find that iWMNs can improve their performance by using the adaptive network coding technique, since the number of wireless transmissions in the network is considerably reduced, and, consequently, i) the medium access contention decreases, ii) the probability of errors in the medium is reduced, and iii) the capacity of the network increases.

show abstract