The use of smart fishing devices (SFDs) is expanding for applications in marine science and fisheries management. These devices signal a capture event through low-power wide-area networks (LPWAN) or non-terrestrial networks (NTN), providing real-time data. A critical challenge in these applications is ensuring the healthy release of non-target species, as post-release stress and survival are strongly species-specific. Consequently, identifying the species that has triggered the device is paramount. This study investigates whether the analysis of an SFD’s GPS periodic position is sufficient to infer the captured species. We examined a dataset comprising 11,905 GPS positions from 983 distinct capture events. Our findings indicate that GPS data is not sufficient for accurate species classification. However, the data can effectively be used to detect self-release events, achieving 72% accuracy from a sequence of 3 GPS positions and 80% accuracy from 12 positions. This finding demonstrates that SFD positional data is a promising tool for remotely monitoring gear availability. Furthermore, the inability to classify species with this data highlights the need to incorporate more granular sensors, such as accelerometers or higher-frequency GPS, in future work.

The Border Gateway Protocol (BGP) is in charge of the route exchange at the Internet scale. Anomalies in BGP can have several causes (misconfigurations, outages and attacks). Forged AS paths are small scale and subtle attacks on BGP and therefore are hard to detect. In this paper we use a Machine Learning (ML) model applying a Long Short-Term Memory (LSTM) based Recurrent Neural Network (RNN) architecture to take the temporal aspect of BGP into account. We show that our ML model is able to detect forged AS path anomalies with an accuracy of 67% and a precision of 72%. These preliminary results outperform the existing proposals and allow us to think that ML on temporal graphs is worth investigating.

The Border Gateway Protocol (BGP) builds the communication routes at the Internet scale. Anomalies in BGP have several causes and can impact the Internet stability. BGP data traces are complex and require specific methods such as machine learning to be processed for anomaly detection. Two types of features are used to study large scale events with machine learning models: graph features or statistical features. Despite the recent interest for the concept of Graph Neural Network (GNN), there is no proposal that adapts GNN for BGP anomaly detection directly from the BGP graph. In this paper, we propose BGNN, a GNN model which detects if a node is involved in a large scale BGP anomaly. Our results show a maximum accuracy of 96% and the model can detect an anomaly after 6 minutes with 90% accuracy. These results are promising and suggest GNN for BGP anomaly detection are worth investigating,

The Border Gateway Protocol (BGP) is in charge of the route exchange at the Internet scale. Anomalies in BGP can have several causes (mis-configuration, outage and attacks). These anomalies are classified into large or small scale anomalies. Machine learning models are used to analyze and detect anomalies from the complex data extracted from BGP behavior. Two types of data representation can be used inside the machine learning models: a graph representation of the network (graph features) or a statistical computation on the data (statistical features). In this paper, we evaluate and compare the accuracy of machine learning models using graph features and statistical features on both large and small scale BGP anomalies. We show that statistical features have better accuracy for large scale anomalies, and graph features increase the detection accuracy by 15% for small scale anomalies and are well suited for BGP small scale anomaly detection.

This paper proposes "LD-Tetrys" (Late Decoding Tetrys), a solution based on an on-the-fly erasure code that attempts to solve the problem of late decoded packets usually considered as lost by the video decoder. LD-Tetrys has the following advantages: i) it drastically improves the trade-off between throughput and quality without modifying the codecs or adding complexity at the encoder side ii) it allows a simple but robust configuration. The only cost is a minor modification of the decoding process and a slight increase in the video decoding complexity. Last but not least, LD-Tetrys requires a much smaller playout buffer to obtain the same perceived video quality, bringing benefits for interactive applications.

Performance of transport protocols on lossy links is a well researched topic, however there are only a few proposals which make use of the opportunities presented by the use of erasure coding within the transport layer, particularly in the multipath transport protocol context. In this paper, we investigate the improvements in the performance of multipath SCTP transport protocol brought by the novel integration of the on-the-fly erasure code into the congestion control and reliability mechanisms of CMT-SCTP. Our contributions include: integration of transport protocol and erasure codes with regards to congestion control and reliability mechanisms; proposal for a variable retransmission delay parameter in the sender (aRTX) adjustment; evaluation of the performance of CMT-SCTP with erasure coding with simulations. Our results show that we can achieve from 10% to 80% improvements in application goodput under lossy multipath network conditions without a significant penalty i.e. with a minimal (10%) overhead due to the encoding-decoding process. We further evaluate the performance of video streaming using an equivalent of partially reliable CMT-SCTP with erasure coding and again demonstrate a solid performance improvement for our proposal, compared to CMT-SCTP. Finally, we provide an analytical evaluation of CMT-SCTP with erasure codes and validate this with simulation results.

Video streaming for heterogeneous types of devices, where nodes have different devices characteristics in terms of computational capacity and display, is usually handled by encoding the video with different qualities. This is not well suited for Peer-To-Peer (P2P) systems, as a single peer group can only share content of the same quality, thus limiting the peer group size and efficiency. To address this problem, several existing works propose the use of Multiple Descriptions Coding (MDC). The concept of this type of video codec is to split a video in a number of descriptions which can be used on their own, or aggregated to improve the global quality of the video. Unfortunately existing MDC codes are not flexible, as the video is split in a defined number of descriptions. In this paper, we focus on the practical feasibility of using a Fountain MDC code with properties similar to existing Fountain erasure codes, including the ability to create any number of descriptions when needed (on the fly). We perform simulations using selected pictures to assess the feasibility of using these codes, knowing that they should improve the availability of the video pieces in a P2P system and hence the video streaming quality. We observe that, although this idea seems promising, the evaluated benefits, demonstrated by the PSNR values, are limited when used in a real P2P video streaming system.

In this paper, we investigate the ability of using Delay Tolerant Networking (DTN) in Public Safety networks, where bandwidth and storage are constrained. We formalize the problem as a Wardrop equilibrium over a time-discretized graph. Driven by our findings, we propose RECOR, a centralized REsource-Constrained ORacle-based DTN routing mechanism, which spreads the demand across multiple store-carry forward paths to satisfy the node storage and link transport constraints observed in intervention situations. By applying the proposed mechanism to real Bluetooth-based DTN traces, we show that the transmission bottleneck can be compensated, but only up to a certain extent, by increasing storage capacity and delay. We also analyze the benefit of strategies that provide more resources to highly connected nodes (e.g. ambulances and firetrucks) which can then feed incentives and policies for DTN network engineering. Finally, the idea presented here is general and suggests the necessity of trace-driven simulation and specific modeling tools for appropriate design of future DTN resource management policies.

We analyze the dynamics of a mobility data set collected in a pipelined disruption-tolerant network (DTN), a particular class of intermittently-connected wireless networks characterized by a 1-D topology. First, we collected and investigated traces of contact times among thousands of participants of a rollerblading tour in Paris. The data set shows extreme dynamics in the mobility pattern of a large number of nodes. Most strikingly, fluctuations in the motion of the rollerbladers cause a typical accordion phenomenon—the topology expands and shrinks with time, thus influencing connection times and opportunities between participants. Second, we show through an analytical model that the accordion phenomenon, through the variation of the average node degree, has a major impact on the performance of epidemic dissemination. Finally, we test epidemic dissemination and other existing forwarding schemes on our traces, and conclude that routing should adapt to the varying, though predictable, nature of the network. To this end, we propose DA-SW (Density-Aware Spray-and-Wait), a measurement-oriented variant of the spray-and-wait algorithm that tunes, in a dynamic fashion, the number of a message copies to be disseminated in the network. The particularity of DA-SW is that it relies on a set of abaci that represents the three phases of the accordion phenomenon: aggregation, expansion, and stabilization. We show that DA-SW leads to performance results that are close to the best case (obtained with an oracle).

This paper introduces a robust point-to-point transmission scheme: Tetrys, that relies on a novel on-the-fly erasure coding concept which reduces the delay for recovering lost data at the receiver side. In current erasure coding schemes, the packets that are not rebuilt at the receiver side are either lost or delayed by at least one RTT before transmission to the application. The present contribution aims at demonstrating that Tetrys coding scheme can fill the gap between real-time applications requirements and full reliability. Indeed, we show that in several cases, Tetrys can recover lost packets below one RTT over lossy and best-effort networks. We also show that Tetrys allows to enable full reliability without delay compromise and as a result: significantly improves the performance of time constrained applications. For instance, our evaluations present that video-conferencing applications obtain a PSNR gain up to 7 dB compared to classic block-based erasure codes.

Delay Tolerant Networks (DTN) do not provide any end to end connectivity guarantee. Thus, transporting data over such networks is a tough challenge as most of Internet applications assume a form of persistent end to end connection. While research in DTN has mainly addressed the problem of routing in various mobility contexts with the aim to improve bundle delay delivery and data delivery ratio, little attention has been paid to applications. This paper investigates the support of streaming-like applications over DTN. We identify how DTN characteristics impact on the overall performances of these applications and present Tetrys, a transport layer mechanism, which enables robust streaming over DTN. Tetrys is based on an on the fly coding mechanism able to ensure full reliability without retransmission and fast in-order bundle delivery in comparison to classical erasure coding schemes. We evaluate our Tetrys prototype on real DTN connectivity traces captured from the Rollerblading tour in Paris. Simulations show that on average, Tetrys clearly outperforms all other reliability schemes in terms of bundles delivery service.

We analyze the dynamics of a mobility dataset collected in a pipelined disruption-tolerant network (DTN), a particular class of intermittently-connected wireless networks characterized by a one-dimensional topology. First, we collected and investigated traces of contact times among a thousand participants of a rollerblading tour in Paris. The dataset shows extreme dynamics in the mobility pattern of a large number of nodes. Most strikingly, fluctuations in the motion of the rollerbladers cause a typical accordion phenomenon - the topology expands and shrinks with time, thus influencing connection times and opportunities between participants. Second, we show through an analytical model that the accordion phenomenon, through the variation of the average node degree, has a major impact on the performance of epidemic dissemination. Finally, we test epidemic dissemination and other existing forwarding schemes on our traces, and argue that routing should adapt to the varying, though predictable, nature of the network. To this end, we propose DA-SW (density-aware spray-and-wait), a measurement-oriented variant of the spray-and-wait algorithm that tunes, in a dynamic fashion, the number of a message copies disseminated in the network. We show that DA-SW leads to performance results that are close to the best case (obtained with an oracle).

Although ironically it does not offer any real-time guarantee, Internet is a popular solution to support multimedia time-constrained applications (e.g. VoIP, Video Conferencing, ...).