Satellite systems typically use physical and link layer reliability schemes to compensate the significant channel impairments, especially for the link between a satellite and a mobile end-user. These schemes have been introduced at the price of an increase in the end-to-end delay, high jitter or out-of-order packets. This is show to have a negative impact both on multimedia and best-effort traffic, decreasing the Quality of Experience (QoE) of users. In this paper, we propose to solve this issue by scheduling data transmission as a function of the channel condition. We first investigate existing scheduling mechanisms and analyze their performance for two kinds of traffic: VoIP and best-effort. In the case of VoIP traffic, the objective is to lower both latency and jitter, which are the most important metrics to achieve a consistent VoIP service. We select the best candidate among several schedulers and propose a novel algorithm specifically designed to carry VoIP over LEO constellations. We then investigate the performance of the scheduling policies on Internet-browsing traffic carried by TCP, where the goal is now the maximize the users' goodput, and select the best candidate in this case.

Satellite transmissions can suffer from high channel impairments, especially on the link between a satellite and a mobile end user. To cope with these errors, physical and link-layer reliability schemes have been introduced at the price of an increased end-to-end delay seen by the transport layer [for example, Transmission Control Protocol (TCP)]. By default, the TCP enables delayed acknowledgment (DelAck) that might increase the end-to-end delay when performing over satellite link-layer recovery schemes. As a matter of fact, even if this option enables decreasing the feedback path load and the stack processing overhead, it might be counterproductive in a satellite context. This motivates the present paper, which aims to quantify the impact of such a TCP option in the context of low Earth orbit satellite constellations. Several simulation measurements are performed with two well-deployed TCP variants, and it is shown that DelAck should be disabled when used over link-layer Hybrid Automatic Repeat Request schemes, particularly when these schemes enable reordering the buffer.

This paper investigates strategies to carry out delay tolerant services over LEO satellite constellations for mobile receiver. In this context, LEO constellations are characterized by important delay variations where propagation impairments are mostly localized on the Land Mobile Satellite (LMS) channel (i.e. on the last hop). To cope with this issue, distinct reliability schemes can be introduced at the physical or link layers. Although their capacity to cope with transmission errors has been demonstrated, these recovery schemes may induce a high jitter that could severely damage TCP's internal timers and reliability schemes. As a matter of fact, transport and link layers' reliability schemes exhibit a clear discrepancy. Following temporal traces representing the delay between a mobile terminal and the last hop satellite from a LEO constellation, we assess how HARQ mechanisms impact on the RTO based retransmission and the duplicate acknowledgments of TCP. Based on ns-2 simulations, we propose a layer-2 buffer that let both link and transport layers to conjointly perform. Our evaluations show an end-to-end data rate increase and more generally illustrate the benefit of reordering packets at the link layer when link-layer erasure coding recovery mechanisms are used conjointly with TCP.

Satellite transmissions can suffer from high channel impairments, especially on the link between a satellite and a mobile end-user. To cope with these errors, physical and link layer reliability schemes have been introduced at the price of an end-to-end delay increase resulting in high jitter. Unfortunately, both the delay and the jitter negatively impacts on multimedia traffic. As a matter of fact, not taking into account the channel state greatly decreases the Quality of Experience (QoE) of VoIP users. In this paper, we propose to solve this issue by scheduling data transmission as a function of the channel condition. We first investigate existing scheduling mechanisms and analyze their performance for VoIP traffic with the objective to lower both latency and jitter, which are the most important metrics to achieve a consistent VoIP service. We select the best candidate among several schedulers and propose a novel algorithm specifically designed to carry VoIP over LEO constellations. Our simulations show that in some scenarios, we double the QoE of VoIP users.