The Musical Metaverse relies on distributed audio, visual, inference, and delivery functions deployed across heterogeneous edge-cloud resources. Queue-aware orchestration frameworks such as SPARQ jointly optimize service placement, routing, and resource allocation, but fix the execution regime of each resource in advance: it is either shared, favoring statistical multiplexing, or guaranteed, favoring performance isolation. This paper introduces MusiSPARQ, a subscription-aware control framework that promotes resource isolation from a fixed infrastructure attribute to a runtime decision. Each switchable resource is represented through coupled shared- and guaranteed-resource pools drawing from the same physical-capacity budget, and at each orchestration epoch the controller selects which class-specific service flows remain shared and which receive isolated capacity. The per-epoch decision jointly accounts for class-specific latency objectives, per-class maximum-latency limits, queueing stability, isolation cost, and switching overhead. In a dynamic virtual-concert scenario, MusiSPARQ reduces the Premium latency-target violation rate from 16.19% to 0.29% relative to fully shared execution, while reserving on average only 0.71% of capacity and, unlike static isolation, keeping Standard and Basic traffic within their latency targets.
Ruifeng Zheng, Zhihan Xu, Pit Hofmann, Frank H. P. Fitzek
This paper investigates whether low-order packet-loss statistics suffice to assess survival-time risk under bursty packet losses, i.e., the probability that no update succeeds for a prescribed number of consecutive slots. Conventional models, such as the Gilbert model, characterize packet losses using the loss rate and mean burst length, but channels sharing these statistics may exhibit substantially different failure risks. We characterize the Markov channels one order above a given low-order fit that share its statistics and analyze their risks under acknowledgment-based scheduling with transmission budgets. We show that, with dense transmission attempts, risk is primarily governed by loss persistence beyond the modeled memory, and the worst-case channel need not be the most persistent one. Under limited budgets, retry spacing makes loss relapse probabilities important and can reverse the risk ordering of compatible channels. Numerical and measurement results show that low-order models can substantially underestimate the risk, highlighting the need for scheduling-dependent channel statistics.
We consider the problem of graph private information retrieval (PIR) in the quantum setting. We focus on a specific type of graphs, namely the forest graphs. A forest graph is a union of disjoint trees. We consider the quantum setting where all the servers share an entangled state. Each server is allowed to send only one qudit per retrieval round. We design a scheme for the forest graphs that achieves the rate of $\frac{3}{N}$. The scheme uses a GHZ state as the initial entangled state between the servers prior to the scheme initiation. We compare our scheme with the only work done in the literature that considered the quantum setting of graph PIR. We show that our scheme is more efficient in this setting. We derive the quantum capacity for the case of $G=3K_2$, which is equal to $4/6$. $G=3K_2$ is a special case of the forest graph. We design a scheme that is specialized for this graph, achieving a rate equal to $4/6$ and provide the corresponding matching upper bound, establishing it as the capacity.
Rubi Debnath, Daniel Bujosa Mateu, Luxi Zhao, Silviu S. Craciunas, Paul Pop, Sebastian Steinhorst
We present TSNBench, the first benchmark for evaluating large language model (LLM) proficiency in Time-Sensitive Networking (TSN), a suite of IEEE 802.1 standards for deterministic communication with bounded latency in safety-critical domains such as autonomous vehicles, aviation, defense, and industrial automation. While LLMs have been extensively evaluated on general knowledge tasks, their capabilities in safety-critical networking domains remain largely unexplored. TSNBench comprises 939 expert-validated multiple-choice questions (MCQs) covering diverse TSN mechanisms, along with 100 open-ended Worst-Case Delay (WCD) computation tasks for each of Credit-Based Shaper (CBS) and Cyclic Queuing and Forwarding (CQF), across varying topologies and traffic flows. The ground-truth (GT) WCD values are computed using a verified Network Calculus (NC) solver for CBS and closed-form upper bounds for CQF. We evaluate 16 LLMs and find that although models achieve 67 to 95% accuracy on MCQs, they fail substantially on open-ended WCD computation. For CBS, the best-performing model, GPT-5, deviates from the GT by 150.2 $μ$s on average (mean absolute error, MAE), and its predicted WCD is within 50 $μ$s of the GT for only 39.4% of flows, while all other models exceed an MAE of 250 $μ$s. Providing the WCD equations in the prompt lowers the error for 9 of 10 models on CQF but only 2 of 9 on CBS. Such errors are large relative to TSN latency budgets and can lead to violations of real-time constraints and unsafe configurations. TSNBench demonstrates that MCQ benchmarks may overestimate LLM capabilities in safety-critical networking.
Sotiris Chatzimiltis, Mahdi Boloursaz Mashhadi, Mohammad Shojafar, Rahim Tafazolli
Next-generation open radio access network (Open RAN) and AI-RAN deployments turn the RAN into a disaggregated, compute-aware platform whose service continuity depends on many control loops acting at different timescales. Resilience in this setting is best seen as a lifecycle spanning anticipation, absorption, adaptation, recovery, and evolution. Yet, existing mechanisms remain stage-specific and are rarely coordinated across the control hierarchy. This paper argues that agentic AI offers an orchestration layer that coordinates this lifecycle, turning operator intent into coordinated action while keeping slow language-model reasoning off the fast control path. Using mass-attachment signaling storms as a running study, we build a discrete-event Open RAN attach simulator that captures retry amplification, shared-compute contention, and non-instantaneous provisioning, and we propose a three-tier framework in which an orchestrator at the service management and orchestration layer sets a standing policy, a per-site non-real-time agent judges storms and anticipates scheduled crowds, and a deterministic near-real-time loop enforces every decision. Across experiments, the agentic controller approaches the resilience of a fully provisioned pool of servers at a quarter of its capacity by reading a scheduled event from a calendar and provisioning ahead of it. No reactive weight setting recovers this once the surge arrives. Free-text operator intents ground into control directives at 94% exact match on previously unseen intents, and a single instruction reshapes live behavior end to end. We also quantify the latency, cost, and contention limits that bound it. The supplementary material provides analytical proof of the derived Theorem on anticipation horizon, as well as the subsequent Corollary results on reactive provisioning limits.
Technology-mediated audience participation (TMAP) enables audiences to influence live performances, yet balancing interaction with meaningful agency remains challenging. While AI-generated music and MR enable new forms of audio and visual audience participation, little is known about how these participation modalities influence audience experiences during live musical performance. In particular, it remains unclear how they shape perceived agency, collaboration, and engagement, and how they complement one another within TMAP systems. We present CrowdForge, an AI- and MR-mediated paradigm where audiences collectively control generative music and virtual environments during a live performance featuring a guitarist and DJ. Through a mixedmethods concert study evaluating three interaction conditions (audio, visual, and combined audio-visual), we demonstrate that observability is the primary driver of audience agency. While visual interaction provided easily observable feedback, audio parameters offered macro-structural control over the music. Audiences tended to use the control enabled by AI audio generation to test musical boundaries, introducing a productive adversarial tension between the crowd and the musicians. This dynamic pushed creative limits, deepened engagement, and generated moments of musical coherence when performers successfully resolved audience-induced challenges. We argue future TMAP systems should prioritize human-centered, highly observable interactions to build agency, while combining the expressiveness of music with the clarity of visual feedback.
Federated learning (FL) is well suited to Internet of Things (IoT) sensing. However, devices carry different sensors and come online in periodic patterns, and as such, the sensors among the participants change from round to round. Under naive averaging, devices without a sensor overwrite its parameters, and a single-sensor device never learns how its sensor relates to the others. Existing methods either redesign local training for mismatched sensors (assuming always-available devices) or select clients without regard to their sensors. We propose Tier-aware Intra-pod Decoupled Encoders (TIDE). Specifically, each sensor has its own encoder and auxiliary head, whose parameters are merged only within the "pod" (i.e., the devices that hold the sensor), while the fusion layer and the task head remain shared by all devices. On three multimodal activity recognition datasets against eight baselines, TIDE reaches 78.06% on ActionSense, 7.63 points above the next method, is the most accurate on all three datasets at half its own communication, and at equal communication leads on ActionSense and is ahead within seed variation on the other two.
Ran Ben-Basat, William Kuszmaul, Michael Mitzenmacher, Amit Portnoy, Shay Vargaftik
Uniform random rotations (URRs) are a common preprocessing step in modern quantization approaches used for gradient compression, inference acceleration, KV-cache compression, model weight quantization, and approximate nearest-neighbor search in vector databases. In practice, URRs are often replaced by randomized Hadamard transforms (RHTs), which preserve orthogonality while admitting fast implementations. The remaining issue is the performance for worst-case inputs. With a URR, each coordinate is distributed as a shifted beta distribution, which converges to a Gaussian distribution in high dimensions. Generally, one RHT is not suitable in the worst case, as individual coordinates can be far from these distributions. We show that after composing two RHTs on any $d$-sized input vector, the marginal distribution of every fixed coordinate of the normalized rotated vector is within $O(d^{-1/2})$ of a standard Gaussian both in Kolmogorov distance and in $1$-Wasserstein distance. We then plug these bounds into the analyses of modern compression schemes, namely DRIVE and QUIC-FL, and show that two RHTs achieve performance that asymptotically matches URRs. However, while two RHTs suffice for scalar quantization, they may be insufficient for Vector Quantization (VQ), which depends on the joint distribution across fixed-size blocks of coordinates (as opposed to only marginal distribution convergence for single coordinates). We prove that a composition of three RHTs leads to decaying coordinate covariance. This ensures that any fixed, bounded, multi-dimensional VQ codebook optimized for URRs has the same expected error when using three RHTs, up to an additive term that vanishes with the dimension. Finally, because practical inputs are rarely adversarial, we propose a linear-time $O(d)$ check on the input's moments to dynamically adapt the number of RHTs used at runtime to improve performance.
Boris Radovič, Istemi Ekin Akkus, Marco Canini, Ruichuan Chen, Alice Dethise, Ivica Rimac, Veljko Pejović
Training models over data distributed across edge devices presents a trade-off: centralizing data simplifies training but raises privacy and regulatory concerns, while full local training may exceed the devices' compute and memory capacity. Split learning addresses this challenge by partitioning the model between clients and the server, allowing part of the computation to be offloaded. However, existing approaches typically rely on static split configurations or simplified assumptions that fail to capture the heterogeneity and complex dynamics of real deployments. We present DynSL, a framework that dynamically allocates model layers between clients and the server by continuously monitoring system metrics and periodically solving an optimization problem that accounts for client and server computation and communication. When beneficial, the system reconfigures the split, enabling the offloading strategy to adapt to changing device capabilities and system conditions. Extensive experiments on a testbed of 24 heterogeneous edge devices show that DynSL reduces training time by up to $\sim\!$30% over the strongest adaptive baseline and by more than 50% over a suboptimal uniform split, with no loss -- and sometimes a gain -- in accuracy.
Liang Song, Jiayue Jin, Peng Sun, Di Li, Yang Liu, Athanassios Skodras, Dimitris Ampeliotis, Yong Yuan 等 12 位作者
Recent advances in Artificial Intelligence (AI), particularly large-scale deep learning, have been driven by monolithic or offline training on massive curated datasets. While successful, these approaches face structural limitations: escalating data and computational demands, limited online adaptability, and a lack of system-level intelligence in open, non-stationary environments. This paper presents Networked AI - a networked intelligent system composed of interacting, heterogeneous, and continuously evolving AI agents. We establish a hierarchical layered architecture comprising an Agent Layer for heterogeneous capability construction, an Interaction Layer for signal-processing-guided online data extraction, and a Regulation Layer for stability-constrained collective evolution. Foundations in data science, signal processing, and system control are hereby synthesized into a unified theoretical framework of Networked AI, providing superior data efficiency, intelligence efficiency, interpretability, and scalability. Theoretical and experimental analysis demonstrates that Networked AI can approach near-linear scaling in resource demand, in contrast to the exponential increase in standalone end-to-end systems. It provides not only online adaptation to out-of-distribution environmental novelties, but also the autonomous evolution towards application-specific intelligence. Application frontiers are discussed for IoT spaces, embodied, spatial, and industrial intelligences, as well as the cellular infrastructure and beyond. Key open challenges are identified towards realizing locally autonomous, collectively intelligent, continuously evolving, globally regulated Networked AI.
Mingxin Li, Enge Song, Yueshang Zuo, Xiaodong Liu, Rong Wen, Qiang Fu, Gianni Antichi, Jian He 等 30 位作者
LLM agents increasingly rely on tool calling to complete tasks. Model Context Protocol (MCP) standardizes tool calling between LLM agents and tool providers. It reduces integration cost and drives rapid growth in tool scale. In production deployments, tool access becomes difficult to scale. On the tool provider side, legacy services are not directly callable through MCP; the rapid protocol development creates ongoing compatibility cost. On the agent side, the number of mounted tools is limited by the LLM context window and inference overhead; mounting a large tool set increases token usage and inference latency and can reduce task success rate. Moreover, for stateful MCP backends with multiple replicas, preserving session affinity increases client-side complexity. We present a cloud-scale MCP Gateway system. It breaks the direct-connect model on the data plane and offloads legacy service integration, consolidating incompatible MCP variants, access control, tool recommendation, and session-aware routing at the gateway. The conversion module supports 315 OpenAPI services in the cloud and integrates MCP transport variants. Session-aware routing decouples the scalability of the gateway layer and the MCP service layer. The hybrid retrieval engine sustains $\sim$98\% Top-15 recall; it scales agent tool access to 3{,}000+ tools with improved tool selection accuracy. We also share the lessons learned from deploying the gateway system in production.
Over-the-Air Federated Learning (OTA-FL) improves communication efficiency by exploiting waveform superposition over wireless Multiple-Access Channels (MAC). However, client heterogeneity, fading, energy constraints, and participation imbalance can degrade both fairness and aggregation accuracy. Existing Pareto-oriented fair OTA-FL schemes may still suffer from large aggregation distortion when highly weighted clients experience poor channel conditions. This paper proposes DisC, a game-theoretic distributed client selection framework for fairness-aware, channel-aware, and energy-efficient OTA-FL. Each client selects its participation level by maximizing a utility that captures Pareto fairness importance, learning relevance, OTA distortion, energy consumption, and delay. We show that DisC admits a Nash equilibrium and can be formulated as an exact potential game, where stable equilibria correspond to locally optimal OTA-FL operating points. DisC therefore offers a distributed, incentive-compatible alternative to centralized client selection under wireless and resource constraints.
Yuma Kaminaga, Hoang D. Le, Cuong T. Nguyen, Anh T. Pham
The emergence of sixth-generation (6G) wireless networks is driving the need for secure group key generation mechanisms resilient to quantum-computing attacks. Conventional quantum key distribution (QKD) protocols are designed for point-to-point communication and are therefore not directly applicable to group key establishment. To address this limitation, this paper proposes a scalable High-Altitude Platform (HAP)-assisted multipartite QKD network architecture that integrates $N$-GHZ-state protocols with a Time Division Multiple Access (TDMA) scheme over free-space optical (FSO) channels. The performance of the proposed system is evaluated using a Qiskit-based simulation framework that incorporates realistic physical-layer impairments of FSO links. Simulation results demonstrate that the proposed architecture significantly improves the group secret key rate (SKR) compared to conventional bipartite approaches, while effectively reducing overall network execution delay. In addition, the analysis reveals the impact of channel impairments and provides practical design insights, such as optimal beam-waist control. The proposed framework demonstrates the potential of HAP-assisted multipartite QKD as an effective security solution for future large-scale 6G wireless networks supporting secure group communications.
Routing-driven timely data collection in Underwater Acoustic Sensor Networks (UASNs) is crucial for marine environmental monitoring, disaster warning, and underwater resource exploration, etc. However, harsh underwater conditions, including high delays, limited bandwidth, and dynamic topologies, make efficient routing decisions challenging in UASNs. In this paper, we propose a smart interrupted routing scheme for UASNs to address dynamic underwater challenges. We first model underwater noise influences from real underwater routing features, e.g., turbulence and storms. We then propose a Software-Defined Networking (SDN)-based Interrupted Software-defined UASNs Reinforcement Learning (ISURL) framework, which ensures adaptive routing through dynamical failure handling (e.g., energy depletion of sensor nodes or link instability) and real-time interrupted recovery. Based on ISURL, we propose the MA-MAPPO algorithm, integrating multi-head attention mask mechanism with MAPPO to filter out infeasible actions and streamline training. Furthermore, to support interrupted data routing in UASNs, we introduce MA-MAPPO\_i, MA-MAPPO with interrupted policy, to enable smart interrupted routing decisions in UASNs. The evaluations demonstrate that our proposed routing scheme achieves exact underwater data routing decisions with faster convergence speed and lower routing delays than existing approaches.
The evolution of 6th generation (6G) networks increasingly relies on satellite-based Non-Terrestrial Networks (NTNs) to extend broadband connectivity to remote and unserved regions, and to support public safety. In this paper we compare two representative and conceptually different satellite constellation architectures, namely Starlink and IRIS 2. Starlink is a commercial private Internet constellation by SpaceX, based on dense Low Earth Orbit (LEO) satellites. It is primarily designed to deliver high-capacity broadband services for civil applications, with performance targets comparable to those of terrestrial networks. In contrast, IRIS 2 is a planned public initiative to be deployed by the European Union, based on a multi-layer combination of LEO, Medium Earth Orbit (MEO), and Geo-stationary Earth Orbit (GEO) satellites. It is primarily designed to provide a secure, resilient, and sovereign infrastructure for government and critical communications. After describing the main technical characteristics of Starlink and IRIS 2, we run a comprehensive simulation campaign to evaluate the design tradeoffs between the two. Specifically, we evaluate the per-cell and per-user achievable capacity, the impact of satellite mobility and handover, and identify the capability of each architecture to support global and reliable connectivity. We also provide design suggestions for possible future IRIS 2 deployment extensions.
Konstantinos Bountrogiannis, Anthony Ephremides, Panagiotis Tsakalides, George Tzagkarakis
This work introduces a framework for analyzing the Age of Incorrect Information (AoII) in a real-time monitoring system with a generic discrete-time Markov source. We study a noisy communication system employing a hybrid automatic repeat request (HARQ) protocol, subject to a transmission rate constraint. The optimization problem is formulated as a constrained Markov decision process (CMDP), and it is shown that there exists an optimal policy that is a randomized mixture of two stationary policies. To overcome the intractability of computing the optimal stationary policies, we develop a multiple-threshold policy class where thresholds depend on the source, the receiver, and the packet count. By establishing a Markov renewal structure induced by threshold policies, we derive closed-form expressions for the long-term average AoII and transmission rate. The proposed policy is constructed via a relative value iteration algorithm that leverages the threshold structure to skip computations, combined with a bisection search to satisfy the rate constraint. To accommodate scenarios requiring lower computational complexity, we adapt the same technique to produce a simpler single-threshold policy that trades optimality for efficiency. Numerical experiments exhibit that both thresholdbased policies outperform periodic scheduling, with the multiplethreshold approach matching the performance of the globally optimal policy.
This paper investigates a reconfigurable intelligent surface (RIS)-assisted rate-splitting multiple access (RSMA) downlink with imperfect eavesdropper channel state information and finite-resolution RIS phases. The common stream carries public information, while user-specific private streams carry confidential information. For a conservative security assessment, each eavesdropper is allowed to remove the public stream and is given the other private streams as side information. We formulate a robust design of t he transmit precoders, isotropicartificial noise, common-rate allocations, and RIS phases that balances confidential and public throughput under rate and power constraints. Bounded errors in the direct and cascaded eavesdropper channels yield an effective-channel uncertainty bound valid across RIS phase updates. To address the resulting nonconvex problem, we construct conservative leakage bounds and convex inner transmit updates, initialize through a feasibility procedure, and search finite artificial-noise and RIS-phase candidates with explicit feasibility checks. For a fixed artificialnoise power and RIS configuration, the feasible inner iterates have a nondecreasing objective that converges in value, and their successive changes vanish. Accepted outer updates remain feasible and are finite in number. Our code is available at https://github.com/ABM-ADAM/RIS RSMA-secure.
Radio-frequency (RF) data synthesis predicts the received signal given transmitter and receiver positions, and is essential for wireless applications. Recent 3D Gaussian Splatting (3DGS)-based methods achieve efficient synthesis at any transmitter location within a surveyed scene, but require a separate model for each receiver. We therefore present RxGS, a shared 3DGS model that first learns receiver-independent Gaussian geometry and then freezes it while learning directional radiance conditioned on receiver position. Its global conditioning branch captures receiver-dependent effects shared across the scene, while a local branch models per-scatterer variations from the receiver's geometry and occlusion. The shared geometry also supports batched rendering through a multi-receiver rasterizer. Across RF datasets, RxGS predicts at receiver positions excluded from training within the surveyed scene while significantly reducing training, inference, and storage costs. Project page: https://kangyangg.com/projects/rxgs/.
Keyvan Aghababaiyan, Baldomero CollPerales, Javier Gozalvez
Future 6 G networks are envisioned as a network of networks (NoN) ecosystem, integrating communication and computing resources across multiple domains. At the deep edge, IoT and end-user devices will form subnetworks for local communication and distributed task processing. These subnetworks will seamlessly integrate into the NoN ecosystem, creating an IoT-edge-cloud continuum. The unified resources across this continuum facilitate dynamic and scalable task offloading, unlocking new possibilities to support emerging services, including critical vertical services with stringent reliability and deterministic service level requirements. In this context, this paper demonstrates that a deterministic approach to task offloading and resource (communication and computing) allocation in the IoT-edge-cloud continuum not only ensures deterministic service levels but also enhances scalability compared to existing task offloading and resource allocation methods. By flexibly managing task completion deadlines while maintaining deterministic (i.e. bounded latency) service levels, deterministic policies achieve a more balanced workload and resource distribution across the continuum, ultimately improving scalability.
The same radio interactions that complicate wireless communication can reveal what is happening around a network. Integrated sensing and communication (ISAC) turns these interactions into opportunities to observe unmanned aerial vehicles (UAVs), road vehicles, and industrial mobile robots. Yet a strong echo is not necessarily a reliable object observation: stationary background can conceal motion, reflections can create virtual objects, and one viewpoint can leave part of an object's motion unobserved. This tutorial explains how known cellular signals are used to estimate propagation, suppress stationary background, interpret reflected paths, and combine observations from different sites. UAV examples under a Third Generation Partnership Project (3GPP) stochastic channel model and ray tracing in a San Francisco urban scene illustrate the mechanisms through spatial layouts, channel profiles, and before/after range--Doppler maps. Local checks reduce false outputs, with some loss of detection. Cooperation improves position accuracy for UAVs detected by both a single site and cooperating sites. The examples show how signal resources, measurement reporting, and application needs shape sensing performance, providing practical guidance for sensing services in complex environments.
Teams in the University Rover Challenge (URC) drive their rovers remotely and often out of sight. The organizers name communication failure as the most common reason a team falls short. A rover therefore needs bandwidth for several camera views and a command path that survives the loss of one radio. We describe and field test the communication stack of the BRACU Mongol-Tori rover, which keeps vision and control on separate radios and bands. Video and RTK corrections travel over a 5.2 GHz bridge with a 2.4 GHz backup, and the Babel routing protocol chooses between them. Every command goes out on two sub-gigahertz radios at once, and the rover uses the backup copy only when the primary falls silent. We tested the backhaul on the bench with a power cut and injected packet loss. We also drove the rover more than a kilometer from the base, and line of sight ended about halfway. On the bench, traffic reached the backup within a few seconds. A metric offset kept it on a primary that lost one packet in ten. In the field, the primary failed at 771 m and the backup carried video and corrections to just over a kilometer. The drive also exposed a limit of Babel, which judges a link only by lost Hello messages. It kept traffic on a primary with a faulty receiver, although the backup reported twice the capacity, and it flapped at the edge of range. We close with lessons on antenna aiming, link metrics and logging.
We examine bounded retained history through a frozen predictive pilot and a subsequent allocation study on a curated UNSW-NB15 flow mirror. Both use an artificial release-order replay based on raw ltime ticks, with equal-tick records querying a shared prior-group state. In the pilot, the model with added retained relational features had lower reconnaissance recall than the current-flow-plus-counters baseline (43.66% versus 47.84%) and lower benign false-positive rate (3.63% versus 6.68%) at separately frozen operating thresholds, together with lower AUROC. This mixed result did not establish a predictive advantage. A development-only audit subsequently found that capacity eviction dominated expiration in the 256-record, 300-tick FIFO state, with a median completed record lifetime of 11 ticks. A protocol specified before its single execution, but after the pilot and post-hoc mechanism audit, compared five label-blind retention arms on the already-used development trace. Its primary equal-source endpoint averaged retained/reference ratios for service-expanded source out-degree and source-destination service diversity. Source-balanced retention at 256 records exceeded FIFO-256 by a pooled endpoint difference of 0.440738 and met the prespecified materiality and subblock rule. Crucially, the event-weighted ratios for both quantities decreased relative to FIFO-256: the result is specific to the equal-source weighting objective and does not establish uniformly improved retention. The study supplies neither evidence of detection improvement nor an explanation of the pilot result. Reused data, concentrated reconnaissance support, proxy timing, and the absence of a qualified independent corpus bound the interpretation.
Hamid Hassani, Anudeep Karnam, Kishor Joshi, Sonia M. Heemstra, Ignas G. Niemegeers, George Exarchakos
In mobile networks, timely and accurate detection of the onset of queueing at path bottlenecks remains challenging for end-to-end congestion control algorithms (CCAs), where end hosts have limited visibility into queue status. Traditional congestion signals, such as packet loss, round-trip time (RTT), and one-way delay (OWD), have well-known limitations for early queue buildup detection. This paper presents QueueTrack, a practical end-host framework to detect the onset of queueing in mobile networks. QueueTrack computes a queue buildup indicator, denoted QBI, from endpoint-observable packet-level timing statistics, without requiring tight synchronization between end hosts or modifications to the existing network infrastructure. We develop and assess the complete QueueTrack machine learning (ML) pipeline, including data collection, labeling, model training, and testing, on a 5G testbed built on the OpenAirInterface platform, under a range of scenarios with competing TCP traffic. The experimental results show that QueueTrack can reliably detect the onset of queueing at the 5G base station using low-complexity classification models such as logistic regression. For example, QueueTrack achieves an F1 score of approximately 80% in a challenging mixed TCP/UDP scenario. These findings confirm that QueueTrack can serve as a practical framework for developing CCAs that target low delay, as well as for network telemetry and network monitoring in mobile networks.
Hamid Hassani, Staji Wilson, Kishor C. Joshi, Sonia M. Heemstra, Ignas G. Niemegeers, George Exarchakos
Emerging delay-sensitive interactive services, such as extended reality and remote operation, demand advanced congestion control algorithms (CCAs) capable of maintaining high data rates while keeping latency low in mobile networks. Existing CCAs typically rely on congestion signals such as packet loss, delay, or explicit congestion notification (ECN) marks, which can be delayed, imprecise, or require extensive upgrades to the network infrastructure. To address these limitations, physical resource block (PRB) utilization has recently been proposed as a reliable early congestion signal in mobile networks [1]. This paper introduces PRISM, a reinforcement-learning-based CCA that uses peak PRB utilization to regulate the send rate under challenging mobility and fading conditions. NS3 simulation results show that, in a mobility scenario where two PRISM flows coexist with two TCP Cubic flows and one TCP BBR flow, PRISM maintains a per-packet mean one-way delay (OWD) of 6.3ms and a $p_{99}$ OWD of 44.6ms, with negligible (0.02%) packet loss, over a 20min high-contention interval. However, over the same interval, the coexisting TCP flows experience 60x higher mean OWDs of 371-390ms and $p_{99}$ OWDs of 563-578ms, with 0.1% packet loss. PRISM achieves a 31x improvement in goodput-over-latency performance, indicating a favorable tradeoff for delay-sensitive applications.
WebAssembly (Wasm) runs untrusted code in a portable, lightweight sandbox and is increasingly used to host services on shared servers and edge platforms. Running the QUIC transport protocol inside Wasm would let each application choose and update its own transport without changing the host. QUIC, however, encrypts every packet, and Wasm offers neither dedicated AES instructions nor carry-less multiplication; a Wasm QUIC stack has been reported to reach only 17.6% of native goodput, with cryptography as its largest cost. Whether the 128-bit SIMD extension of Wasm (WASM-SIMD) can recover this cost, and what limits a pure-Wasm stack once it does, has remained unclear. Here we show that explicit WASM-SIMD kernels recover much of the cryptographic cost, after which non-cryptographic costs become the limit. On two hosts connected by 10-Gigabit Ethernet, kernels that process eight AES blocks or four ChaCha20 blocks per pass raise receiver goodput over a non-SIMD baseline by 61.34% for AES-128-GCM and 31.55% for ChaCha20-Poly1305, whereas compiler auto-vectorization gains only 1.28% for AES-128-GCM. ChaCha20-Poly1305 still delivers 1.99 times the goodput of AES-128-GCM, confirming end to end the preference that stand-alone cipher benchmarks suggest. Receiver-side profiling combined with Amdahl's law predicts that even native-speed cryptography would leave Wasm at roughly one third of the goodput of the fastest native stack. Cryptographic acceleration is therefore necessary but not sufficient: memory allocation, runtime overhead and per-datagram UDP I/O now bound pure-Wasm transports. We anticipate that this per-category cost breakdown will guide further optimization of Wasm network stacks and the design of system interfaces for batched I/O.
Yuanyi Zhu, Zijian Li, Xin Ai, Zixuan Chen, Sen Liu, Yang Xu
Network simulation plays a crucial role in both networking research and industry. Existing commonly-used Discrete Event Simulations (DES) are based on the callback mechanism for discrete events. However, as network functions grow in complexity, callbacks reveal a fundamental structural limitation: a callback can only represent a single point in a blocking operation and cannot capture the full operation as a coherent whole. This forces developers to manually decompose sequential network logic into scattered callback fragments, leading to stack ripping, callback hell, and heavy development and debugging burdens. We propose CoDES (\textbf{Co}routine-based \textbf{DES}), a coroutine-based programming paradigm for network simulation that addresses this structural limitation. CoDES targets development productivity rather than execution performance. Realizing CoDES in existing DES simulators is non-trivial, as coroutine suspension must be coordinated with the event scheduler, introducing challenges in coroutine resource conflicts, operation-level computation requirements, compatibility with existing callback-based simulators, and overhead. We implement a CoDES-based framework in ns-3, address all four challenges, and release it as open source. We evaluate CoDES by implementing Message Passing Interface (MPI), High Precision Congestion Control (HPCC), and Routing Information Protocol (RIP) in ns-3, spanning the transport, congestion control, and routing layers. Compared with callback-based counterparts, CoDES reduces lines of code by up to 51.4% and nesting depth by up to 50%, with no accuracy loss and negligible runtime and memory overhead.
Farhad Mehran, Charles Turyagyenda, Dritan Kaleshi
Open Radio Access Networks (RANs) aim to transform the existing wireless network architectures by adopting open interfaces, disaggregating and virtualising RAN components, and enabling dynamic control of the system via data-driven methods. To accelerate the development of interoperable RAN solutions, rigorous assessment of Open Fronthaul and E2 interfaces based on O-RAN Alliance Interoperability Testing (IOT) specifications plays a key role as the integration of these interfaces seems to be very challenging, e.g. due to the need for alignment of Fronthaul IOT profiles and unforeseen challenges for E2 interface integration as Near-Real-Time RAN Intelligent Controller (Near-RT RIC) is relatively a new component introduced within the RAN architecture. In this paper, selected parts of the latest findings and lessons learned from interoperability assessment of Open Fronthaul and E2 interfaces for 13 multi-vendor Open RAN systems deployed in SmartRAN Open Network Interoperability Centre (SONIC) Labs are presented.
We enable backscatter communication in the battery-less mobile Internet of Things (IoT). Backscatter communication is extensively studied in static settings. Existing designs are, however, fundamentally mismatched with mobility and time-varying energy patterns. Channel conditions rapidly fluctuate, impacting the achievable data rates and thus transmission costs. Energy availability varies unpredictably, possibly forcing devices to remain quiescent to recharge energy buffers. The two issues compound each other: while recharging, a battery-less mobile IoT device may miss more favorable channel conditions. We design a lightweight decision system that dynamically determines when to transmit by checking short-term trends in signal strength, while using Non-volatile Memory (NVM) to retain packets in unfavorable channel conditions and across energy failures. Using a prototype we built and real-world mobility and power traces, we compare our design against a rate-adaptive baseline that only considers the instantaneous channel conditions. Experimental results show that our system improves throughput by up to 5.16x while reducing transmission energy consumption by up to 47.3%, with only 0.23% - 7.3% additional energy overhead.
This paper proposes a spatial-temporal learning-based routing framework for dynamic low Earth orbit (LEO) satellite networks. The routing problem is formulated as a partially observable Markov decision process (POMDP), where each satellite independently selects its next hop using network-state observations and destination information. A graph attention network (GAT) extracts spatial features from the dynamic satellite topology and queue states, while a long short-term memory (LSTM) network captures their temporal evolution. The resulting representation is integrated into a deep Q-network (DQN) for destination-aware next-hop selection without a separate traffic-prediction stage. Simulations in a 45-satellite LEO constellation under normal and bursty traffic demonstrate packet delivery performance comparable to a graph neural network (GNN)-enhanced routing baseline, although the proposed method incurs higher delay and queue occupancy. Ablation studies further demonstrate the complementary benefits of spatial and temporal learning.
As network traffic becomes increasingly dynamic and bursty, conventional centralized traffic engineering (TE) suffers from long control-loop delays caused by network-state aggregation, centralized computation, and policy distribution. In this paper, we propose MATE, a multi-agent TE method for proactive distributed scheduling in hybrid networks with partial deployment of Segment Routing over IPv6 (SRv6). MATE uses historical traffic stability to characterize the burst risk of each traffic demand and path sensitivity to quantify the ability of candidate paths to accommodate additional traffic, thereby enabling differentiated proactive scheduling tailored to heterogeneous traffic characteristics. MATE adopts an offline-centralized training and online-distributed execution architecture. During training, a shared global critic evaluates joint actions under a reward combining maximum link utilization (MLU) with variability-weighted path sensitivity, while counterfactual difference rewards provide agent-specific credit for actor learning. During online execution, edge agents generate traffic-splitting policies from local observations and apply them through SRv6, avoiding centralized optimization in each scheduling interval. Experiments on the China Environment for Network Innovations (CENI) testbed and simulations on the Sanet topology show that MATE outperforms a multi-agent reinforcement learning (MARL) ablation baseline without agent-specific credit assignment, COPE, and TEAL in most evaluated scenarios and approaches centralized FIGRET on measured traffic traces. On CENI-Abilene, MATE achieves MLU within 10% of the linear-programming optimum in up to 90.1% of test instances. MATE also achieves lower median normalized MLU than FIGRET in most tested failure scenarios, with mean neural-inference times of 2.05--3.14~ms.