On elastic traffic via contention resolution diversity slotted aloha satellite access

Abstract

Summary This paper presents a performance study relative to the coupling of the Transmission Control Protocol (TCP) with the Contention Resolution Diversity slotted aloha (CRDSA) protocol, in the case of greedy TCP connections (also called elephants) on Digital Video Broadcasting-Return Channel via a geostationary satellite. CRDSA, which takes advantage of interference cancellation algorithms for collision/contention resolution, has already exhibited interesting performance when the power levels of all received bursts are perfectly balanced. In this paper, we extend the study to a more realistic case, where a certain spreading of the bursts' power levels is taken into account. The consequent capture effect even facilitates the collision resolution mechanism and yields an improvement in the overall TCP performance with respect to the balanced case. Furthermore, in certain conditions, the adoption of packet level forward error correction allows achieving even higher peaks of throughput than the expected ones

Small satellites and CubeSats: Survey of structures, architectures, and protocols

“…There are two main families of MAC protocols for satellite networks: scheduled access, such as the combined free/demand assignment multiple access (CFDAMA), based on resource reservation mechanisms and scheduled transmissions in order to avoid channel contention, and random access, such as contention resolution diversity slotted ALOHA (CRDSA), which provides contention resolution mechanisms and, for some of them, interference cancellation techniques. Research efforts have been performed to increase their reliability and efficiency (eg, previous studies), but their implementation in small satellites needs further evaluation to consider the system constraints. –Upper layers: Considering a CubeSat network as a multihop network where satellites are not always in contact with ground stations or among them, there could not be always persistent paths between data sources and destinations, because satellite links are not always up. In addition, in case all data generated by CubeSat sensors are processed and stored in a control station linked to the ground stations through wired cable, in the path between CubeSat and control station, there are different kinds of links (satellite and terrestrial), and the communication through each link could be based on different protocols.…”

Application protocols enabling internet of remote things via random access satellite channels

“…The completion time in a emulated cellular network is used as key performance indicator, and, according to the authors, OPCbased communications outperform CoAP and MQTT-based data transfer, at the price of a larger overhead. MQTT is TCPbased, and the contributions provided in [9], [10], [11], [12] shed some lights on how TCP behaves in presence of a random access satellite link dominated by collisions, because it may represent a limiting factor on the achievable throughput in satellite environments. On the other side, CoAP is UDP-based, and a comparison is on order when dealing with random access satellite links, in order to provide some reference figures on the achievable performance level.…”

M2M Traffic via Random Access Satellite links: Interactions between Transport and MAC Layers

“…Furthermore, the TCP performance over a satellite Diversity Slotted Aloha (DSA) [20] RA medium is analyzed for a single TCP connection in [21]; the authors assume that the MAC layer operates at a target load G * and that the Network Control Center (NCC) periodically broadcast an activity factor (i.e., a throttling parameter) to dynamically adjust the proposed flow control policy applied at Return Channel Satellite Terminals (RCSTs), in order to maximize the TCP throughput. Analogously, [22] analyzes TCP throughput over CRDSA when packet-level Forward Error Correction (FEC) techniques are employed; instead, in this work, we do not use any packet-level FEC techniques. Furthermore, a single TCP connection is modeled through a fluidic model in [22], less precise than the packet level simulator in use in this work.…”

“…Analogously, [22] analyzes TCP throughput over CRDSA when packet-level Forward Error Correction (FEC) techniques are employed; instead, in this work, we do not use any packet-level FEC techniques. Furthermore, a single TCP connection is modeled through a fluidic model in [22], less precise than the packet level simulator in use in this work. Finally, reference [23] considers a similar scenario, i.e.…”